Abstract
Objectives
The purpose of the study was to examine potential differences between two approaches to defining adolescent weight misperception. Specifically, weight status perception was compared with self-reported weight status and actual weight status (based on body mass index percentiles calculated from self-reported and actual weights and heights, respectively). Furthermore, the accuracy of assigning weight status based on body mass index percentiles calculated from self-reported weights and heights was assessed by comparing them with actual weight status.
Methods
Data were extracted from Team Up for Healthy Living, an 8-week, school-based obesity prevention program in southern Appalachia. Participants (N = 1509) were predominately white (93.4%) and ninth graders (89.5%), with approximately equivalent representation of both sexes (50.7% boys).
Results
The study revealed significant differences between the approaches to defining weight misperception (χ2 = 16.2; P = 0.0003).
Conclusions
Researchers should interpret study findings with awareness of potential differences based on the method of calculating weight misperception.
Keywords: obesity, overweight, weight misperception
Approximately 34% of 12- to 19-year-olds are overweight or obese.1 Numerous physical and psychosocial health consequences2,3 and economic costs4 have directed attention to the need for effective prevention/intervention programs. Adolescents who are overweight or obese are significantly more likely than their normal-weight counterparts to misperceive their weight status.5 This is concerning because overweight or obese adolescents who underestimate their weight status have been found to be less motivated to change their weight-related health behaviors.6 Conversely, normal-weight adolescents who overestimate their weight status tend to engage in more harmful weight-related practices7; therefore, future studies examining the potential effects of weight misperception across weight categories, including underweight, are warranted.
The literature comparing the use of objective/actual versus subjective/self-reported height and weight measurements in the calculation of body mass index (BMI) is growing. A systematic review8 of 64 studies examining the relation between these two methods of assessment suggests trends of underreporting for weight/BMI and overreporting of height, with some differences based on sex and populations studied. Another literature review9 echoes the results of self-reported data underestimating overweight prevalence and suggests that self-reported data should be used only if these are the sole source of data. These issues are further complicated in the context of defining weight misperception.
To date, the most common methodological approach to estimating weight misperception has involved calculating discrepancy scores between perceived weight status (assessed via a single-item question or body/figure silhouettes) and weight status calculated with self-reported6,10–13 or actual weights and heights.14–17 To our knowledge, no studies have directly compared utilization of different methods to define weight misperception, which could affect findings. Other studies have used additional strategies for defining weight misperception such as comparison of differences between self-reported weight versus actual weight18 and perceived weight versus weight status measured by waist circumference.19
The purpose of this study was to examine potential differences between the two most common approaches to defining weight misperception, underestimation and overestimation. Specifically, weight status perception was compared with self-reported weight status and actual weight status based on BMI percentiles calculated from self-reported and actual weights and heights, respectively. Furthermore, the accuracy of assigning weight status based on BMI percentiles calculated from self-reported weights and heights was assessed by comparing it with actual weight status.
Methods
Procedures
Data were extracted from waves 1 and 2 of Team Up for Healthy Living, an 8-week, school-based obesity prevention program in southern Appalachia. Participating schools were randomized to either treatment or control sites. Baseline data were collected from the 1509 adolescents who were enrolled in the program in January–February 2012 and August–September 2012. Trained staff administered a comprehensive survey, which was distributed in Lifetime Wellness classes and assessed among other characteristics, lifestyle-related behaviors, and demographic characteristics. In addition, anthropometric measurements were completed as described below. Written consent was obtained from participants before enrollment in this institutional review board–approved study. The details of the design and methods of the study have been described in detail elsewhere.20,21
Measures
Demographics
The demographic characteristics included in the analysis were age, sex, race/ethnicity, grade level, parents' highest level of education, and family income. Date of birth (month/day/year) was subtracted from date of measurement to calculate students' age in years and months. Sex was self-reported via a question asking, “What is your gender?” Race/ethnicity was classified as American Indian/Alaska Native, Asian, black or African American, Hispanic or Latino, Native Hawaiian or Other Pacific Islander, non-Hispanic white or non–African American, or Other by asking, “How do you describe yourself? (Select one or more responses).” Grade level was obtained by asking, “What grade are you in?” Students were asked to report their mother's/father's highest level of education, with response options that ranged from less than high school to college degree and family household income (options ranged from <$20,000 to ≥$75,000).
Perceived Weight Status
Weight perception was assessed by one question: How do you describe your weight? Response options included “very underweight,” “slightly underweight,” “about the right weight,” “slightly overweight,” and “very overweight.” Given the small percentage of students who described themselves as very underweight (1.01%) or very overweight (6.12%), the responses were collapsed into the adjacent response, resulting in three categories consistent with the Centers for Disease Control and Prevention (CDC) nomenclature: underweight (ie, “very underweight” and “slightly underweight”), healthy weight (ie, “about the right weight”), and overweight/obese (ie, “slightly overweight” and “very overweight”).
Self-Reported Weight Status
Self-reported weight and height were assessed by asking students, “How much do you weigh?” (pounds) and “How tall are you?” (feet and inches). To obtain BMI percentile scores using weight in kilograms and height in centimeters, we used the following formula for the transformation: weight in kg = weight in lb × 0.45 and height in cm = height in feet × 30.48 + height in in. × 2.54. We then estimated age- and sex-specific BMI percentile scores based on the CDC 2000 growth charts.22 In addition, students were categorized as underweight (<5th percentile), healthy weight (≥5th–<85th percentile), and overweight/obese (≥85th percentile). For self-reported weight, there were 5 participants <31 lb and 1 participant >1900 lb. For self-reported height, there were 2 participants whose height was <4 ft (reported as only 2 and 3 ft, respectively) and 3 participants >7 ft tall (reported as 8, 12, and 53 ft, respectively). In addition, there was 1 participant who self-reported height as <0 in. (reported as −1 in.) and 4 participants whose height was >12 in. (13, 15, 42, and 69 inches, respectively). These extreme values were set as missing. According to cutoffs defined in the 2000 CDC growth charts,22 biologically implausible values (BIVs) were calculated. After calculating BIV, BMI-related variables (eg, BMI percentile) were set with missing if BIV was >0 (n = 6).
Actual Weight Status
Trained staff collected each student's weight and height using calibrated scales and portable stadiometers. Weight was measured to the nearest 0.1 kg and height to the nearest 0.1 cm. Students were asked to remove heavy outer garments and shoes. We used the same method as described above to obtain percentile scores and assign weight status categories. Once again, BIVs were calculated and BMI-related variables were set to missing if BIV was >0 (n = 11).
Weight Misperception
We calculated weight misperception by comparing the three weight status perception category responses with the self-reported and actual weight status categories. Adolescents were classified as underestimating their weight when their perceived weight category was lower than their self-reported or actual weight category (eg, perceived weight category was underweight and self-reported/actual weight category was healthy weight). Adolescents were classified as overestimating their weight when their perceived weight category was higher than their self-reported or actual weight category (eg, perceived weight category was overweight and self-reported/actual weight category was underweight). Finally, adolescents were classified as accurately estimating their weight category when their perceived weight category matched their self-reported or actual weight category (eg, perceived weight category was healthy weight and self-reported/actual weight category was healthy weight). A similar approach was taken to compare weight status based on self-reported as compared with actual weights and heights.
Data Analysis
Participants' characteristics in the analytic sample (n = 1243) were compared with those not included (n = 266) because of incomplete self-reported weight and height data to assess for a potential selection bias. χ2 tests were used to determine significance for categorical variables (expressed in frequencies with percentage values) and independent t tests were used to determine the significance for continuous variables (expressed as mean ± standard deviation). To satisfy the aim of our study, weight misperception was calculated and compared using two approaches: using self-reported or actual weight and height. For the percentage data of multiple comparisons in categorical data analysis, the correction of adjustment of the significant level was the number of pairwise comparison minus one.23 All of the analyses were performed using SAS version 9.2 (SAS Institute, Cary, NC).
Results
Participants
Participants (N = 1509) were predominately white (93.4%) and ninth graders (89.5%) with approximately equivalent representation of both sexes (50.7% male). A majority of parents indicated a high school or greater education, with 27.1% of mothers and 20.9% of fathers possessing a college degree (Table 1).
Table 1. Sample characteristics.
| Characteristics | Overalla | Boys | Girls |
|---|---|---|---|
| Age, y, mean (SD) | 14.9 (0.7) | 14.9 (0.8) | 14.8 (0.7) |
| Sex, n (%) | |||
| Female | 744 (49.3) | NA | 744 (100) |
| Male | 765 (50.7) | 765 (100) | NA |
| Grade in school, n (%) | |||
| 9th | 1309 (89.5) | 655 (88.6) | 654 (90.3) |
| 10th | 94 (6.4) | 53 (7.2) | 41 (5.7) |
| 11th | 37 (2.5) | 21 (2.8) | 16 (2.2) |
| 12th | 23 (1.6) | 10 (1.4) | 13 (1.8) |
| Race/ethnicity, n (%) | |||
| American Indian or Alaska Native | 14 (1.0) | 8 (1.1) | 6 (0.8) |
| Asian | 4 (0.3) | 3 (0.4) | 1 (0.1) |
| Black or African American | 11 (0.8) | 6 (0.8) | 5 (0.7) |
| Hispanic or Latino | 39 (2.7) | 20 (2.7) | 19 (2.7) |
| Native Hawaiian or Other Pacific Islander | 1 (0.1) | 0 (0) | 1 (0.1) |
| White non-Hispanic or non–African American | 1364 (93.4) | 692 (92.9) | 672 (93.9) |
| Other | 28 (1.9) | 16 (2.2) | 12 (1.7) |
| Highest level of education: mother, n (%) | |||
| Less than high school | 86 (6.0) | 38 (5.3) | 48 (6.8) |
| High school graduate or GED | 416 (29.1) | 217 (30.1) | 199 (28.1) |
| Some college | 308 (21.5) | 153 (21.2) | 155 (21.9) |
| College degree | 387 (27.1) | 193 (26.8) | 194 (27.4) |
| Unknown | 233 (16.3) | 120 (16.6) | 113 (15.9) |
| Highest level of education: father, n (%) | |||
| Less than high school | 121 (8.5) | 59 (8.2) | 62 (8.8) |
| High school graduate or GED | 474 (33.3) | 245 (34.1) | 229 (32.5) |
| Some college | 217 (15.2) | 121 (16.8) | 96 (13.6) |
| College degree | 298 (20.9) | 152 (21.1) | 146 (20.7) |
| Unknown | 314 (22.1) | 142 (19.8) | 172 (24.4) |
| Family household income, n (%) | |||
| <$20,000 | 56 (3.9) | 30 (4.1) | 26 (3.6) |
| $20,000–$44,999 | 111 (7.6) | 71 (9.7) | 40 (5.5) |
| $45,000–$74,999 | 112 (7.7) | 65 (8.9) | 47 (6.5) |
| ≥$75,000 | 128 (8.8) | 84 (11.5) | 44 (6.1) |
| Unknown | 1048 (72.0) | 481 (65.8) | 567 (78.3) |
| Perceived weight status (%) | |||
| Underweight | 169 (11.4) | 106 (14.1) | 63 (8.6) |
| Healthy weight | 808 (54.3) | 419 (55.6) | 389 (53.0) |
| Overweight/obese | 511 (34.3) | 229 (30.4) | 282 (38.4) |
| Self-reported weight status (%) | |||
| Underweight | 23 (1.9) | 13 (2.1) | 10 (1.6) |
| Healthy weight | 737 (59.3) | 342 (54.6) | 395 (64.1) |
| Overweight/obese | 483 (38.9) | 272 (43.4) | 211 (34.3) |
| Actual zBMI, mean (SD) | 0.9 (1.1) | 1.0 (1.1) | 0.8 (1.0) |
| Actual weight status (%) | |||
| Underweight | 17 (1.1) | 11 (1.5) | 6 (0.8) |
| Healthy weight | 782 (52.5) | 363 (48.1) | 419 (56.9) |
| Overweight/obese | 692 (46.4) | 381 (50.5) | 311 (42.3) |
Percentages in each column were adjusted to total approximately 100%. Perceived weight status included collapsing or fitting “very underweight” and “slightly underweight” into underweight, “about the right weight” into healthy weight, and “slightly overweight” and “very overweight” into overweight/obese categories. Self-reported and actual weight status categories were assigned via age- and sex-specific BMI percentile scores based on the CDC 2000 growth charts. BMI, body mass index; CDC, Centers for Disease Control and Prevention; GED, General Educational Development; SD, standard deviation; zBMI, age- and sex-standardized BMI.
Missing data for the overall sample was as follows: grade (n =46), race/ethnicity (n =48), mothers' education (n =79), fathers' education (n =85), family household income (n =54), perceived weight status (n =21), self-reported weight status (n =266), actual zBMI (n =7), and actual weight status (n =18).
Weight Misperception Assigned via Self-Reported Weight Status Versus Actual Weight Status
Our study revealed significant differences between the two approaches when calculating weight misperception for the overall sample, χ2 = 16.20, P = 0.0003; male subjects, χ2 = 10.94, P = 0.0042; and female subjects, χ2 = 8.06, P = 0.018. As shown in Table 2, using self-reported weight status (compared with using actual weight status) resulted in a smaller percentage of underestimates (21.9% vs 26.7%, 29.9% vs 34.7%, and 13.8% vs 18.6%, respectively, for the overall sample, male and female subjects) and greater percentage of overestimates (7.4% vs 4.6%, 4.7% vs 1.9%, and 10.3% vs 7.3%, respectively, for the overall sample, male and female subjects).
Table 2. Proportions of weight misperception by methodology.
| Perceived/self-reported | Perceived/actual | χ2 | Nominal Pa |
Actual Pb |
|
|---|---|---|---|---|---|
| Overall, % | |||||
| 2 × 2 table | 1.18 | NA | 0.278 | ||
| Accurate | 70.7 (n = 874) | 68.7 (n = 1011) | |||
| Misperceptionc | 29.4 (n = 363) | 31.3 (n = 460) | |||
| 3 × 2 table | 16.20 | NA | 0.0003 | ||
| Underestimation | 21.9 (n = 271) | 26.7 (n = 393) | 8.40 | 0.0038 | 0.0076 |
| Accurate | 70.7 (n = 874) | 68.7 (n = 1011) | 1.18 | 0.178 | 0.356 |
| Overestimation | 7.4 (n = 92) | 4.6 (n = 67) | 10.10 | 0.0015 | 0.003 |
| Boys, % | |||||
| 2 × 2 table | 0.62 | NA | 0.4308 | ||
| Accurate | 65.5 (n = 408) | 63.4 (n = 472) | |||
| Misperceptionc | 34.5 (n = 215) | 36.6 (n = 272) | |||
| 3 × 2 table | 10.94 | NA | 0.0042 | ||
| Underestimation | 29.9 (n = 186) | 34.7 (n = 258) | 3.59 | 0.058 | 0.116 |
| Accurate | 65.5 (n = 408) | 63.4 (n = 472) | 0.62 | 0.4308 | 0.8616 |
| Overestimation | 4.7 (n = 29) | 1.9 (n = 14) | 8.60 | 0.0034 | 0.0068 |
| Girls, % | |||||
| 2 × 2 table | 0.55 | NA | 0.4598 | ||
| Accurate | 75.9 (n = 466) | 74.1 (n = 539) | |||
| Misperceptionc | 24.1 (n = 148) | 25.9 (n = 188) | |||
| 3 × 2 table | 8.06 | NA | 0.018 | ||
| Underestimation | 13.8 (n = 85) | 18.6 (n = 135) | 5.42 | 0.0199 | 0.0398 |
| Accurate | 75.9 (n = 466) | 74.1 (n = 539) | 0.55 | 0.4598 | 0.9196 |
| Overestimation | 10.3 (n = 63) | 7.3 (n = 53) | 3.72 | 0.0539 | 0.1078 |
P value was obtained from multiple comparisons.
For the percentage data of multiple comparisons in categorical data analysis, the correction of adjustment of the significant level was the number of pairwise comparisons minus one.
Misperception includes underestimation and overestimation.
Table 3 shows sex differences within weight misperception methodologies. Male subjects were significantly more likely than female subjects to misperceive (either underestimate or overestimate) their weight status despite whether self-reported weight status (34.5% vs 24.1%; χ2 = 16.15, P < 0.0001) or actual weight status was used (36.6% vs 25.9%; χ2 = 19.59, P < 0.0001). Female subjects were more likely than male subjects to overestimate their weight status (10.3% vs 4.7%; χ2 = 14.12, P = 0.0004, using self-reported weight status; 7.3% vs 1.9%; χ2 = 24.74, P < 0.0002, using actual weight status). Male subjects were more likely than female subjects to underestimate their weight status (29.8% vs 13.8%; χ2 = 46.34, P < 0.0002, using self-reported weight status; 34.7% vs 18.6%; χ2 = 48.73, P < 0.0002, using actual weight status).
Table 3. Sex differences within weight misperception methodologies.
| Boys | Girls | χ2 | Nominal Pa |
Actual Pb |
|
|---|---|---|---|---|---|
| Perceived/self-reported, % | |||||
| 2 × 2 table | 16.15 | NA | <0.0001 | ||
| Accurate | 65.5 (n = 408) | 75.9 (n = 466) | |||
| Misperceptionc | 34.5 (n = 215) | 24.1 (n = 148) | |||
| 3 × 2 table | 53.99 | NA | <0.0001 | ||
| Underestimation | 29.9 (n = 186) | 13.8 (n = 85) | 46.34 | <0.0001 | <0.0002 |
| Accurate | 65.5 (n = 408) | 75.9 (n = 466) | 16.15 | <0.0001 | <0.0002 |
| Overestimation | 4.7 (n = 29) | 10.3 (n = 63) | 14.12 | 0.0002 | 0.0004 |
| Perceived/actual, % | |||||
| 2 × 2 table | 19.59 | NA | <0.0001 | ||
| Accurate | 63.4 (n = 472) | 74.1 (n = 539) | |||
| Misperceptionc | 36.6 (n = 272) | 25.9 (n = 188) | |||
| 3 × 2 table | 65.45 | NA | <0.0001 | ||
| Underestimation | 34.7 (n = 258) | 18.6 (n = 135) | 48.73 | <0.0001 | <0.0002 |
| Accurate | 63.4 (n = 472) | 74.1 (n = 539) | 19.59 | <0.0001 | <0.0002 |
| Overestimation | 1.9 (n = 14) | 7.3 (n = 53) | 24.74 | <0.0001 | <0.0002 |
P value was obtained from multiple comparisons.
For the percentage data of multiple comparisons in categorical data analysis, the correction of adjustment of the significant level was the number of pairwise comparisons minus one.
Misperception includes underestimation and overestimation.
Weight Status Assigned via Self-Reported Versus Actual Weight and Height Data
When comparing the relation between self-reported versus actual weight status, 87.4% (n = 1077) of students' categories were accurate or matched. Another 10.1% (n = 125) underestimated their actual weight status, whereas 2.5% (n = 31) overestimated their actual weight status. Male and female subjects did not differ significantly regarding percentages of underestimation, accuracy, or matching and overestimation (data not shown). Notably, 266 participants were missing self-reported weight and height data. As shown in Table 4, we examined for potential differences in demographic and variables of interest across the analytic and nonanalytic groups. Of importance, no significant differences emerged in demographic characteristics of participants between the two groups except for family household income.
Table 4. Comparison of participants in the self-reported/actual analytic sample vs participants in nonanalytic sample as a result of incomplete self-reported weight/height data (N = 1509).
| Characteristicsa | Analytic sample (n = 1243) |
In nonanalytic sample (n = 266) |
A | P |
|---|---|---|---|---|
| Age, y, mean (SD) | 14.85 (0.8) | 14.86 (0.8) | 0.870 | |
| Sex, n (%) | 0.671 | |||
| Female | 616 (49.6) | 128 (48.1) | ||
| Male | 627 (50.4) | 138 (51.9) | ||
| Grade, n (%) | 0.133 | |||
| 9th | 1089 (89.6) | 220 (88.7) | ||
| 10th | 79 (6.5) | 15 (6.1) | ||
| 11th | 32 (2.6) | 5 (2.0) | ||
| 12th | 15 (1.2) | 8 (3.2) | ||
| Race/ethnicity, n (%) | 0.233 | |||
| American Indian or Alaska Native | 13 (1.1) | 1 (0.4) | ||
| Asian | 4 (0.3) | 0 (0.0) | ||
| Black or African American | 10 (0.8) | 1 (0.4) | ||
| Hispanic or Latino | 30 (2.5) | 9 (3.6) | ||
| Native Hawaiian or Other Pacific Islander | 1 (0.9) | 0 (0.0) | ||
| White non-Hispanic or non–African American | 1135 (93.7) | 229 (92.0) | ||
| Other | 19 (1.6) | 9 (3.6) | ||
| Highest level of education: mother, n (%) | 0.531 | |||
| Less than high school | 70 (5.9) | 16 (6.6) | ||
| High school graduate or GED | 346 (29.2) | 70 (28.8) | ||
| Some college | 260 (21.9) | 48 (19.8) | ||
| College degree | 326 (27.5) | 61 (25.1) | ||
| Unknown | 185 (15.6) | 48 (19.8) | ||
| Highest level of education: father, n (%) | 0.304 | |||
| Less than high school | 101 (8.5) | 20 (8.4) | ||
| High school graduate or GED | 389 (32.8) | 85 (35.7) | ||
| Some college | 181 (15.3) | 36 (15.1) | ||
| College degree | 260 (21.9) | 38 (16.0) | ||
| Unknown | 255 (21.5) | 59 (24.8) | ||
| Family household income, n (%) | 0.008 | |||
| <$20,000 | 45 (3.7) | 11 (4.5) | ||
| $20,000–$44,999 | 95 (7.9) | 16 (6.5) | ||
| $45,000–$74,999 | 102 (8.4) | 10 (4.1) | ||
| ≥$75,000 | 116 (9.6) | 12 (4.9) | ||
| Unknown | 851 (70.4) | 197 (80.1) | ||
| Perceived weight status (%) | 0.173 | |||
| Underweight | 142 (11.5) | 27 (10.8) | ||
| Healthy weight | 683 (55.2) | 125 (49.8) | ||
| Overweight/obese | 412 (33.3) | 99 (39.4) | ||
| Self-reported weight status (%) | ||||
| Underweight | 23 (1.9) | NA | ||
| Healthy weight | 737 (59.3) | NA | ||
| Overweight/obese | 483 (38.9) | NA | ||
| Actual zBMI, mean (SD) | 0.9 (1.0) | 1.0 (1.2) | 0.128 | |
| Actual weight status (%) | 0.320 | |||
| Underweight | 13 (1.1) | 4 (1.6) | ||
| Healthy weight | 657 (53.3) | 125 (48.5) | ||
| Overweight/obese | 563 (45.7) | 129 (50.0) |
Percentages in each column were adjusted to total approximately 100%. Perceived weight status included collapsing or fitting “very underweight” and “slightly underweight” into underweight, “about the right weight” into healthy weight, and “slightly overweight” and “very overweight” into overweight/obese categories. Self-reported and actual weight status categories were assigned via age- and sex-specific BMI percentile scores based on the CDC 2000 growth charts. BMI, body mass index; CDC, Centers for Disease Control and Prevention; GED, General Educational Development; SD, standard deviation; zBMI, age- and sex-standardized BMI.
Missing data as follows: grade (n = 46), race/ethnicity (n = 48), mothers' education (n = 79), fathers' education (n = 85), family household income (n = 54), perceived weight status (n =21), self-reported weight status (n =266), actual zBMI (n = 7), and actual weight status (n = 18).
Discussion
The present study found significant differences between two approaches to defining weight misperception in early adolescence. Specifically, using self-reported weight and height data as compared with actual weight and height data may result in lower occurrences of underestimation and higher occurrences of overestimation. Consistent with studies that examined weight misperception among adolescents,6,12,13,16,24 we found that male subjects were more likely than female subjects to misperceive their weight status using either methodology. In addition, female subjects were more likely than male subjects to overestimate, whereas male subjects were more likely to underestimate their weight status, findings that are consistent with previous research.10,13,24
Our study found that 87% of students who self-reported weight and height had correspondence between their weight status calculated with self-reported data and weight status calculated with actual data. Beck and colleagues14 reported similar results among a sample of fifth-grade students, with approximately 79% accurately reporting their height or weight. Research among children indicates that older youth estimate their height and weight with better precision than do younger children,14 which may explain the high percentage of correspondence between self-reported and actual weight status in the present study. In an effort to better understand these findings in the context of this study, we examined for differences between students who did provide versus those who did not provide self-reported weight and height data. Buttenheim and colleagues found obesity prevalence to be higher among younger teens who did not self-report weight and height compared with those who did.25 This finding was not replicated in our sample, and only family household income was found to differentiate the groups, with this finding being difficult to interpret because of high numbers reporting “unknown” on this variable.
We further predicted that it is possible that those who provided self-report data were more aware of their weight and height or had recently obtained this information. The state of Tennessee long ago incorporated the Coordinated School Health (CSH; http://www.tennessee.gov/education/schoolhealth/aboutcsh.shtml) approach, including monitoring of student weight and height. AAn informal telephone survey was completed, with several CSH coordinators in schools participating in the Team Up for Health Living project to assess whether students were typically provided weight and height information to explain the high rates of correspondence. These discussions revealed a variety of methods for collecting and reporting weight and height data across school settings; however, the majority of CSH coordinators reported typically not providing these data during actual measurements with students. This was informative, although a more rigorous research design is needed to confirm these findings. The high rate of correspondence may be considered unexpected given the high rate of overweight/obesity in the current sample (approximately 46%) and that findings from a systematic review on the accuracy of proxy measures in assessing adolescent overweight suggest self-reported weight and height result in systematic underestimation of BMI and, consequently, of actual overweight status.9 Future research may seek to explain the relation between weight status calculated with self-reported data as compared with actual weight and height data. Future studies also may examine the potential influence of overweight/obesity rates associated with self-perceived weight status. For example, in areas where there are higher rates of overweight/obesity, there may be more acceptance or less concern about higher weight.
The strengths of the present study include the use of a large sample size from a cluster-randomized study design, including weight and height data collected via both self-report and objective measurements. The limitations of the study include the focus on a specific age group, primarily ninth-grade students, as compared with a broader range of ages. It may be that factors specific to this group, such as sexual maturity, could influence self-perceptions of weight. In addition, we were unable to determine when participants had last measured weight or height, a factor that may bias results.
Although there were no differences in terms of actual weight status among adolescents providing versus not providing self-reported weight and height, there is still the chance that the absence of their internal self-report data may contribute to their inability to make a judgment about themselves when responding to the perceived weight status question. Furthermore, the differences between perceived weight response options and the CDC weight status categories may be problematic. For example, when students were asked about their perceived weight, there was no mention of “health” or “obesity,” as in the CDC weight status categories nomenclature (ie, healthy weight, overweight/obese), and it is unknown how students who reported they were “about the right weight” may have actually perceived their weight status (eg, possibly not as a healthy weight as it would have been assigned), suggesting some difference in semantic correspondence. These concerns are not unique to this study and may be applied to the larger literature using these methodologies.
This study also was conducted in the Appalachian region of the United States, where there are high rates of diabetes and obesity.26 Future studies wishing to replicate these findings should include a larger age range and multiple geographical locations, and may consider a larger ethnic distribution because of suggestions that weight misperception may differ across ethnic groups.6,11–13
Conclusions
Given the emerging literature demonstrating health risks associated with weight misperception, this study suggests that researchers should interpret findings with an awareness of potential differences based on the method of calculating weight misperception. Some large-scale weight misperception studies10,13 use proxy measures such as self-reported weight and height data. It has been suggested that these measures frequently replace actual measures as a result of convenience or cost.8,27 Based on our findings, the results from such studies should be interpreted with caution because these studies may reveal lower occurrences of underestimation and higher occurrences of overestimation than studies using actual weight and height measurements. Instead, researchers should aim to collect actual weight and height from participants in an effort to determine weight misperception. One alternative solution that may improve large-scale research relying on self-report data is a two-method measurement that requires actual data on only a portion of the sample, providing researchers with a strategy to reduce bias and improve estimates.28 The prevalence of weight misperception in the present study further highlights the importance of screening for weight misperception in both female and male subjects. Furthermore, the differences in measurement techniques highlight the need for careful interpretation of findings and may have implications for clinical practice. Specifically, practitioners may benefit from using actual as opposed to self-report weight and height data, especially when assessing weight misperception.
Key Points.
Studies using self-reported weights and heights of adolescents in the estimation of weight misperception may reveal lower occurrences of underestimation and higher occurrences of overestimation than studies using actual weights and heights.
Male subjects were more likely than female subjects to misperceive their weight status despite the use of self-reported weight status or actual weight status in estimating weight misperception.
Researchers should interpret study findings with awareness of the potential differences based on the method of calculating weight misperception.
Acknowledgments
The authors thank Taylor McKeehan, Megan Herring, and Laura Maphis for data management, writing assistance, and proofreading of the manuscript.
This study was funded by the National Institutes of Health (R01MD006200).
Footnotes
Was the telephone survey part of the present study or part of Tennessee's CSH initiative? Unclear. Also, if part of Tennessee's program, a reference citation is required. Or does ref 9 cover all of this material?
Update for ref 20 needed.
The authors have no financial relationships to disclose and no conflicts of interest to report.
This col head needs a title.
References
- 1.Ogden CL, Carroll MD, Kit BK, et al. Prevalence of obesity and trends in body mass index among US children and adolescents, 1999-2010. JAMA. 2012;307:483–490. doi: 10.1001/jama.2012.40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Dietz WH. Health consequences of obesity in youth: childhood predictors of adult disease. Pediatrics. 1998;101:518–525. [PubMed] [Google Scholar]
- 3.Griffiths LJ, Parsons TJ, Hill AJ. Self-esteem and quality of life in obese children and adolescents: a systematic review. Int J Pediatr Obes. 2010;5:282–304. doi: 10.3109/17477160903473697. [DOI] [PubMed] [Google Scholar]
- 4.Wolf AM, Colditz GA. Current estimates of the economic cost of obesity in the United States. Obes Res. 1998;6:97–106. doi: 10.1002/j.1550-8528.1998.tb00322.x. [DOI] [PubMed] [Google Scholar]
- 5.Maximova K, McGrath JJ, Barnett T, et al. Do you see what I see? Weight status misperception and exposure to obesity among children and adolescents. Int J Obes (Lond) 2008;32:1008–1015. doi: 10.1038/ijo.2008.15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Edwards NM, Pettingell S, Borowsky IW. Where perception meets reality: self-perception of weight in overweight adolescents. Pediatrics. 2010;125:e452–e458. doi: 10.1542/peds.2009-0185. [DOI] [PubMed] [Google Scholar]
- 7.Talamayan KS, Springer AE, Kelder SH, et al. Prevalence of overweight misperception and weight control behaviors among normal weight adolescents in the United States. ScientificWorldJournal. 2006;6:365–373. doi: 10.1100/tsw.2006.70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Connor Gorber S, Tremblay M, Moher D, et al. A comparison of direct vs. self-report measures for assessing height, weight and body mass index: a systematic review. Obes Rev. 2007;8:307–326. doi: 10.1111/j.1467-789X.2007.00347.x. [DOI] [PubMed] [Google Scholar]
- 9.Sherry B, Jefferds ME, Grummer-Strawn LM. Accuracy of adolescent self-report of height and weight in assessing overweight status: a literature review. Arch Pediatr Adolesc Med. 2007;161:1154–1161. doi: 10.1001/archpedi.161.12.1154. [DOI] [PubMed] [Google Scholar]
- 10.Eichen DM, Conner BT, Daly BP, et al. Weight perception, substance use, and disordered eating behaviors: comparing normal weight and overweight high-school students. J Youth Adolesc. 2012;41:1–13. doi: 10.1007/s10964-010-9612-8. [DOI] [PubMed] [Google Scholar]
- 11.Foti K, Lowry R. Trends in perceived overweight status among overweight and nonoverweight adolescents. Arch Pediatr Adolesc Med. 2010;164:636–642. doi: 10.1001/archpediatrics.2010.90. [DOI] [PubMed] [Google Scholar]
- 12.Krauss RC, Powell LM, Wada R. Weight misperceptions and racial and ethnic disparities in adolescent female body mass index. J Obes. 2012;2012 doi: 10.1155/2012/205393. 205393. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Park E. Overestimation and underestimation: adolescents' weight perception in comparison to BMI-based weight status and how it varies across socio-demographic factors. J Sch Health. 2011;81:57–64. doi: 10.1111/j.1746-1561.2010.00561.x. [DOI] [PubMed] [Google Scholar]
- 14.Beck J, Schaefer CA, Nace H, et al. Accuracy of self-reported height and weight in children aged 6 to 11 years. Prev Chronic Dis. 2012;9:E119. doi: 10.5888/pcd9.120021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Gray W, Crawford MJ, Follansbee-Junger K, et al. Associations between actual and perceived weight and psychosocial functioning in children: the importance of child perceptions. Child Obes. 2012;8:147–154. doi: 10.1089/chi.2011.0033. [DOI] [PubMed] [Google Scholar]
- 16.Schiefelbein EL, Mirchandani GG, George GC, et al. Association between depressed mood and perceived weight in top and high school age students: Texas 2004–2005. Matern Child Health J. 2012;16:169–176. doi: 10.1007/s10995-010-0733-1. [DOI] [PubMed] [Google Scholar]
- 17.Yan AF, Zhang G, Wang MQ, et al. Weight perception and weight control practice in multiethnic sample of US adolescents. South Med J. 2009;102:354–360. doi: 10.1097/SMJ.0b013e318198720b. [DOI] [PubMed] [Google Scholar]
- 18.Conley A, Boardman JD. Weight overestimation as an indicator of disordered eating behaviors among young women in the United States. Int J Eat Disord. 2007;40:441–445. doi: 10.1002/eat.20383. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Cuypers K, Kvaløy K, Bratberg G, et al. Being normal weight but feeling overweight in adolescence may affect weight development into young adulthood-an 11-year follow up: the HUNT Study, Norway. J Obes. 2012;2012 doi: 10.1155/2012/601872. 601872. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Slawson DL, Dalton WT, III, Wang L, et al. College students as facilitators in reducing adolescent obesity disparity in southern Appalachia: Team Up for Healthy Living. doi: 10.1016/j.cct.2015.04.012. In press.B. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Southerland J, Behringer B, Slawson D. Using the give-get grid to understand potential expectations of engagement in a community-academic partnership. Health Promot Pract. 2013;14:909–917. doi: 10.1177/1524839913477657. [DOI] [PubMed] [Google Scholar]
- 22.Kuczmarski RJ, Ogden CL, Guo SS, et al. 2000 CDC growth charts for the United States: methods and development. Vital and Health Stat. 2002;11246:1–190. [PubMed] [Google Scholar]
- 23.Cao R, Qian J, Chen PY. Multiple comparisons in categorical data analysis. Nan Fang Yi Ke Da Xue Xue Bao. 2010;30:118–120. [PubMed] [Google Scholar]
- 24.Bodenlos JS, Rosal MC, Blake DR, et al. Obesity prevalence, weight-related beliefs, and behaviors among low-income ethnically diverse national Job Corps students. J Health Dispar Res Pract. 2010;3:106–114. [Google Scholar]
- 25.Buttenheim AM, Goldman N, Pebley AR. Underestimation of adolescent obesity. Nurs Res. 2013;62:195–202. doi: 10.1097/NNR.0b013e318286b790. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Centers for Disease Control and Prevention. Estimated county-level prevalence of diabetes and obesity—United States, 2007. MMWR Morb Mortal Wkly Rep. 2009;58:1259–1263. [PubMed] [Google Scholar]
- 27.Elgar FJ, Roberts C, Tudor-Smith C, et al. Validity of self-reported height and weight and predictors of bias in adolescents. J Adolesc Health. 2005;37:371–375. doi: 10.1016/j.jadohealth.2004.07.014. [DOI] [PubMed] [Google Scholar]
- 28.Drake KM, Longacre MR, Dalton MA, et al. Two-method measurement for adolescent obesity epidemiology: reducing the bias in self-report of height and weight. J Adolesc Health. 2013;53:322–327. doi: 10.1016/j.jadohealth.2013.03.026. [DOI] [PMC free article] [PubMed] [Google Scholar]