To examine trends in perceived overweight among US adolescents, including trends in perceived overweight among overweight and nonoverweight adolescents overall and by sex and race/ethnicity.
Trend analyses of serial cross-sectional data.
National Youth Risk Behavior Surveys conducted in 1999, 2001, 2003, 2005, and 2007.
Nationally representative samples of US high school students in each survey year.
Main Outcome Measures
All students with a body mass index at or higher than the 85th percentile were considered “overweight,” while those with a body mass index lower than the 85th percentile were considered “nonoverweight.” Students who perceived themselves as “slightly overweight” or “very overweight” were considered to perceive themselves as overweight.
Among all students and among most subgroups, the prevalence of overweight increased from 1999 to 2007. The prevalence of perceived overweight did not change. Among nonoverweight students, the prevalence of perceived overweight decreased overall, among white males, and among white, black, and Hispanic females. Among overweight students, few trends in the prevalence of perceived overweight were detected; only among overweight black males did the prevalence of perceived overweight increase.
Weight perception is an important predictor of diet and weight management behaviors. Decreases in the prevalence of perceived overweight among nonoverweight students have positive implications for reducing unhealthy weight control behaviors. Among overweight students, interventions are needed to increase their recognition of being overweight because those who do not perceive themselves as overweight are unlikely to engage in weight control practices.
The prevalence of overweight and obesity among adolescents has increased dramatically in recent decades. Between 1970 and 2000, the prevalence of adolescent overweight and obesity (ie, body mass index [BMI] ≥ the 85th percentile of the Centers for Disease Control and Prevention [CDC] growth charts) more than doubled,1 with the prevalence of obesity (ie, BMI ≥ the 95th percentile) increasing more than 3-fold.2 According to recent data from the National Health and Nutrition Examination Survey, approximately 34% of all 12- to 19-year-olds are overweight or obese; approximately 18% are obese.3
Interventions aimed at preventing childhood and adolescent overweight and obesity have primarily focused on the behavioral determinants of overweight and obesity, such as physical inactivity and unhealthy dietary behaviors, though meta-analyses show that behavioral interventions have had limited success.4,5 One frequently overlooked predictor of behavior change is weight perception. Research has established a strong association between self-perceived weight status and nutritional and weight management behaviors. Studies among high school students have shown that perceiving oneself as overweight is a better predictor of dieting or exercising to control weight than actual body weight is.6,7 Previous research has also found that body weight perceptions are often inaccurate compared with BMI calculated from either self-reported or measured height and weight.8 The implication of these findings is that overweight adolescents who do not perceive themselves to be overweight are unlikely to engage in weight control practices.6- 8 Meanwhile, adolescents who are not overweight but perceive themselves as such may be at risk for eating disorders, such as anorexia nervosa, or other unhealthy dieting and exercise practices.8- 10
Studies of weight perception among adolescents have examined single points in time. It is unknown whether adolescents' perceptions of being overweight have changed over time. It is possible that the increase in overweight and obesity has led to greater acceptance of heavier body weights as heavier body weights have become the social norm.8,11 Several studies support this idea. Maximova et al,12 for example, found that higher parent BMI and higher average BMI among schoolmates were associated with greater underestimation of overweight status among adolescents. Another study by Cash and colleagues11 conducted among college students showed that while the average body weight had increased over time, body image among young college women had actually improved between the mid-1990s and 2001. Among adolescents who are not overweight, a reduction in the prevalence of those who perceive themselves as overweight would be an encouraging finding. However, an increase in misperception among adolescents who are overweight may hinder efforts to change behaviors that reduce body weight and limit further weight gain.13
In this context, the purpose of this study was to examine recent trends in the prevalence of overweight and obesity among US adolescents and to determine whether the percentage of adolescents who perceive themselves as overweight has changed over time. Because previous studies have shown differences in BMI3 or weight perception6- 8,14,15 by sex and race/ethnicity, we examined trends among demographic subgroups of students. Additionally, we examined trends in perceived overweight among nonoverweight and overweight adolescents overall and among demographic subgroups of nonoverweight and overweight adolescents.
The Youth Risk Behavior Surveillance System was developed by CDC to monitor the prevalence of priority health risk behaviors among youth over time. The national school-based Youth Risk Behavior Survey has been conducted biennially since 1991. In 1999, the Youth Risk Behavior Survey questionnaire was modified to include questions on self-reported height and weight to calculate respondent BMI. We used data from the 1999, 2001, 2003, 2005, and 2007 Youth Risk Behavior Surveys, all of which included measures of self-reported height and weight.
In each survey year, a similar 3-stage cluster sample design was used to obtain a nationally representative sample of students in grades 9 through 12. Sampling strategies and the psychometric properties of the questionnaire have been reported elsewhere.16- 18
Student participation in the survey was anonymous and voluntary, and local parental permission procedures were followed. The CDC institutional review board granted approval for the national Youth Risk Behavior Survey. Students completed the self-administered questionnaire during a regular class period. Responses were recorded directly on a computer-scannable questionnaire booklet. Missing data were not statistically imputed. Sample size and school, student, and overall response rates were similar across survey years (Table 1).16,19- 22 A weighting factor was applied to each student record to adjust for nonresponse and the oversampling of black and Hispanic students.
Demographic characteristics included in this analysis were sex, race/ethnicity, and grade level (Table 1). Data for students of all races/ethnicities were included in all analyses, but only results for white, black, and Hispanic students are shown; results for students of other races/ethnicities are not reported because of heterogeneity within this subgroup and insufficient sample size to report on American Indian or Alaska Native, Asian, Native Hawaiian or other Pacific Islander, or multiple race (non-Hispanic) subgroups individually.
Body mass index, calculated from self-reported height and weight data as weight in kilograms divided by height in meters squared, was categorized based on age- and sex-specific growth curves and standard definitions from CDC.23 For the purposes of this analysis, all students with a BMI at or higher than the 85th percentile (ie, those who were overweight or obese) were considered “overweight,” while all those with a BMI lower than the 85th percentile were considered “nonoverweight.”
Perceived weight status was measured by the question “How do you describe your weight?” Response options were “very underweight,” “slightly underweight,” “about the right weight,” “slightly overweight,” and “very overweight.” Responses were collapsed into 2 categories; students who described themselves as “slightly overweight” or “very overweight” were considered to perceive themselves as overweight and compared with those who described themselves as “very underweight,” “slightly underweight,” or “about the right weight.”
For analysis of secular trends, a linear time variable was created using orthogonal coefficients that reflected the biennial spacing of the surveys.
To account for the complex sample design of the survey, we conducted all analyses using SUDAAN.24 Tests for differences by demographic subgroup in the prevalence of overweight, perceived overweight among all students, perceived overweight among nonoverweight students, and perceived overweight among overweight students included respondents from all survey years combined. We determined pairwise differences in prevalence estimates by sex, race/ethnicity, and sex and race/ethnicity using t tests. Differences were considered significant at P < .05.
We examined trends in the prevalence of overweight and perceived overweight for all high school students and among demographic subgroups of students. The time variable was entered into logistic regression models that tested for secular trends among the total high school population and among demographic subgroups of students (subgroups by sex, race/ethnicity, and sex and race/ethnicity) while controlling for changes in the demographic characteristics (sex, race/ethnicity, and grade) of the high school student population over time. We then stratified the sample by overweight status and ran similar regression models testing for trends in perceived overweight among nonoverweight and overweight adolescents overall and among demographic subgroups of nonoverweight and overweight adolescents. Regression coefficients (B) for trends were considered statistically significant at P < .05.
Across all survey years, 26.7% of all students were overweight and 29.9% of students perceived themselves as overweight (Table 2). All pairwise subgroup differences tested in the prevalence of overweight and perceived overweight were statistically significant. However, the patterns of subgroup differences were not the same for overweight and perceived overweight. For example, the prevalence of overweight was higher among males than females (30.7% vs 22.6%; P < .001), though the percentage of students who perceived themselves as overweight was higher among females than males (36.0% vs 24.0%; P < .001). Additionally, among all students, male students, and female students, the prevalence of overweight was higher among black youth than white youth, but white youth had a higher prevalence of perceived overweight than black youth.
Results of the trend analyses demonstrated significant linear increases in the prevalence of overweight between 1999 and 2007. Among all students, the prevalence increased from 25.1% to 28.8% (B = 0.026; P < .001). Significant increases were also found among several demographic subgroups. The prevalence of overweight increased among both male and female students when all racial/ethnic groups were combined and among white, black, and Hispanic students when male and female students were combined. Analyses by sex and race/ethnicity showed that the prevalence of overweight increased among white males and among white and black females. However, there were no significant trends in the percentage of students who perceived themselves as overweight between 1999 and 2007, neither overall nor among any of the demographic subgroups.
Overall, 16% of nonoverweight students perceived themselves as overweight (Table 3). Among nonoverweight students, the percentage of female students who perceived themselves as overweight was approximately 3 times greater than the prevalence of male students who perceived themselves as overweight (23.8% vs 7.5%; P < .001). Among males, the percentage of nonoverweight students who perceived themselves as overweight was greatest among Hispanic students (9.1%), followed by white students (7.5%) and black students (4.0%); all pairwise differences were statistically significant. Among females, the prevalence of nonoverweight students who perceived themselves as overweight was significantly higher among white (25.2%) and Hispanic (24.8%) students than among black students (12.2%).
Trend analyses demonstrated significant linear decreases in the prevalence of nonoverweight students who perceived themselves as overweight. Among all students, the percentage of nonoverweight students who perceived themselves as overweight decreased from 17.7% in 1999 to 14.0% in 2007 (B = −0.031; P < .001). Decreases in the percentage of nonoverweight students who perceived themselves as overweight were found among female students when all racial/ethnic groups were combined and among white, black, and Hispanic students when male and female students were combined; analyses by sex and race/ethnicity also showed decreases among white males and among all 3 racial/ethnic subgroups of female students.
Across all survey years, the percentage of overweight students who perceived themselves as such was 68.5% (Table 4). The prevalence of perceived overweight was significantly higher among overweight females than among overweight males (79.5% vs 60.6%; P < .001). Among overweight males, the prevalence of perceived overweight was higher among white (63.8%) and Hispanic (62.3%) students than among black students (44.8%). Among overweight females, the prevalence of perceived overweight was greatest among white females (85.2%), followed by Hispanic females (81.0%) and then black females (64.7%).
Trend analyses showed that, overall, the prevalence of overweight students who perceived themselves as overweight has not changed over time. Only one significant trend was detected among any of the demographic subgroups; among overweight black males, the percentage of students who perceived themselves as overweight increased from 39.7% in 1999 to 48.1% in 2007 (B = 0.042; P = .03).
Our study showed that while the prevalence of overweight and perceived overweight varied by sex and race/ethnicity, there were similar trends in the prevalence of overweight and perceived overweight across subgroups. Specifically, the prevalence of overweight increased linearly among all US high school students and among most of the subgroups examined, but despite the increase in overweight, there was no change in the prevalence of perceived overweight. Our findings with respect to trends in overweight differ from those found in an analysis of National Health and Nutrition Examination Survey data that found no trends in the prevalence of children and adolescents ages 2 through 19 years who had a BMI for age at or higher than the 85th percentile between 1999-2000 and 2005-2006.3 It is possible that these differences in results are attributable to the different study populations or using measured height and weight (National Health and Nutrition Examination Survey) vs self-report (Youth Risk Behavior Surveillance System) to calculate BMI. To our knowledge, no other national studies have examined trends in perceived overweight among adolescents.
Further examination of trends in perceived overweight by overweight status showed that the percentage of nonoverweight students who perceived themselves as overweight decreased between 1999 and 2007 among the total population and among most demographic subgroups. These decreases in perceived overweight may mean that fewer nonoverweight adolescents will engage in unhealthy and extreme weight control behaviors.8- 10 Despite these decreases, there is still much room for improvement among certain subgroups. Specifically, the percentage of students who perceived themselves as overweight remains particularly high among nonoverweight white females (21.6% in 2007) and Hispanic females (22.3% in 2007). It is important to understand the determinants of the recent improvements in weight perception accuracy among nonoverweight students to inform interventions to further reduce the number of nonoverweight students, particularly these subgroups of females, who believe they are overweight and are therefore at risk for eating disorders and engaging in unhealthy weight control practices.
Among overweight students, approximately two-thirds recognize that they are overweight, though this percentage has not changed between 1999 and 2007. As the prevalence of overweight has increased, the one-third of overweight students who do not perceive themselves as such represent a growing number of students. This has important public health implications, as adolescents who are overweight but do not perceive themselves as such are unlikely to engage in weight control practices.6- 8 Overweight adolescents who adopt healthy lifestyle changes to reduce body weight can improve health outcomes both now and later, when they become adults. It is somewhat encouraging that misperception has not increased over time among overweight adolescents, as has been found among overweight adults.25 However, greater awareness of healthy body weight and strategies to improve weight perception accuracy among adolescents are needed to affect behavior change. Correcting misperceptions during obesity prevention programs may promote readiness to change and facilitate the adoption of healthy nutrition and physical activity behaviors, improving the effectiveness of new and existing behavioral interventions.12,26 Further, efforts to improve the recognition of overweight tailored toward certain subgroups may be necessary. Our study showed that while male students were more likely to be overweight than female students and black students were more likely to be overweight than white and Hispanic students, overweight male students and overweight black students were less likely than their respective counterparts to perceive themselves as overweight. In addition, we found that the percentage of overweight black males who perceive themselves as overweight is particularly low; despite an increase since 1999, the percentage of overweight black males who perceived themselves as overweight was just 48.1% in 2007. Improving weight perception accuracy among these subgroups may in turn promote healthy lifestyle changes that can help to reduce obesity-related health disparities.
Extreme and unhealthy weight control behaviors are common among overweight adolescents; additionally, among overweight adolescents body dissatisfaction is a risk factor for engaging in such behaviors.26 Therefore, it is important when trying to increase the recognition of overweight to address unhealthy weight control behaviors, develop the skills needed for positive behavior change, promote body satisfaction, and discuss realistic weight goals.26
There are several limitations to our study. Our study used self-reported height and weight to determine BMI and overweight status. Studies show that adolescents tend to overestimate their height and underestimate their weight.27 Therefore, the prevalence of overweight in our study is likely lower than the true prevalence of overweight among US high school students. Misperceptions of body weight may also be more prevalent than our study has shown as BMI calculated from self-reported height and weight shows better concordance with weight perceptions than does measured BMI.8 As a result, overweight students who are misclassified as nonoverweight in our study also may not have perceived themselves as overweight; if so, this would further support the need to increase the recognition of overweight status. It is unknown whether the tendencies to overestimate height and underestimate weight have changed over time for the total population or differentially among any of the subgroups. Therefore, it is hard to conclude how misclassification of actual weight status might affect the results of the trend analyses conducted herein. Future research should examine trends in weight perception by overweight status using measured height and weight. Another limitation is that we do not have information on other factors that might influence body weight perception, such as exposure to overweight and obesity. Research has shown that adolescents whose parents and classmates have heavier BMIs are more likely to underestimate their own weight status.12 In addition, because the data are cross-sectional, we cannot infer that the rise in overweight and obesity is causal to the trends in perceived overweight status found in this study. However, our findings suggest that as the prevalence of overweight has increased, nonoverweight students have become more aware that they are not in fact overweight, while a substantial proportion overweight students continue not to recognize that they are overweight.
To better understand the trends in perceived weight status observed herein, the determinants of body weight perception among subgroups of adolescents, including those of other racial/ethnic groups not studied herein, should be further researched. Understanding the determinants and frame of reference for weight perceptions (ie, cultural norms, comparisons with personally desired weight, medical standards, peers, media images) may help to inform interventions to help adolescents accurately assess their body weight. Additionally, understanding how the factors influencing adolescent body weight perceptions may have changed over time and continued monitoring of these factors may help inform such interventions. Accurate weight perceptions, and simultaneous promotion of body satisfaction, may help to prevent unhealthy weight control behaviors among adolescents and to promote positive behavior change among overweight adolescents.
Correspondence: Kathryn Foti, MPH, Division of Adolescent and School Health, Centers for Disease Control and Prevention, Mailstop K-33, 4770 Buford Hwy, Atlanta, GA 30341 (email@example.com).
Submitted for Publication: September 23, 2009.
Author Contributions: The authors had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Foti. Analysis and interpretation of data: Foti and Lowry. Drafting of the manuscript: Foti. Critical revision of the manuscript for important intellectual content: Foti and Lowry. Statistical analysis: Foti and Lowry.
Financial Disclosure: None reported.
Funding/Support: Ms Foti is a health scientist in the Division of Adolescent and School Health at CDC. At the time this article was written, Ms Foti was a fellow in the Division of Adolescent and School Health at CDC, sponsored by the Association of Schools of Public Health and CDC. Dr Lowry is a medical officer in the Division of Adolescent and School Health at CDC.
Disclaimer: The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of CDC.
Foti K, Lowry R. Trends in Perceived Overweight Status Among Overweight and Nonoverweight Adolescents. Arch Pediatr Adolesc Med. 2010;164(7):636–642. doi:10.1001/archpediatrics.2010.90