Overall, perceptions of being at “about the right weight” appear advantageous for youth physical and mental health, regardless of BMI classification, whereas perceptions at either extreme (overweight or underweight) may negatively impact health behaviours and mental health. Instead of considering weight misperceptions as problematic, some researchers have proposed that underestimations of weight status may offer resiliency among individuals with overweight or obesity. Promoting “about right” WPs and preventing change to overweight or underweight perceptions may offer an effective public health strategy for supporting youth health over time. However, limited prospective evidence exists on factors that shape perceptions of weight status over time. The current study examined modifiable predictors of one-year change in weight perception among youths. We used 2-year linked data of 18,112 grade 9–12 students from Year 3 (Y3:2014–2015) and Year 4 (Y4:2015–2016) of the COMPASS study. Generalized Estimating Equation models tested screen use, physical activity, and bullying victimization as predictors of change from perceptions of “about the right weight” to “overweight” or “underweight” perceptions, adjusting for Y3 covariates (body mass index, ethnicity, and grade) and school cluster. Results support the value of team sports among females and resistance exercise among males as protective against changes to overweight or underweight perceptions over one year. Also, various forms of bullying victimization predicted overweight perceptions in males and females. Watching TV/movies or messaging/texting for over 2 hours/day was associated with overweight and underweight perceptions, respectively, in females only. Playing video/computer games for over 2 hours/day was associated with overweight perceptions in males and underweight perceptions in females. Findings support the potential of bullying prevention, limiting certain screen use, and supporting engagement in team sports for females and resistance exercise for males as strategies to maintain perceptions of being at “about the right weight.”
1. Introduction
Body image represents a multidimensional construct, encompassing how individuals view, feel, think, and act toward their physical appearance [1]. Weight perception (WP) refers to individuals’ subjective appraisal of their body weight. In the research literature, misperceptions have been said to occur when there is a discrepancy between individuals’ self-perceptions of their weight and their “objective” or measured weight status (as typically defined by body mass index (BMI) classification). WPs have long been studied in the eating disorders field and, more recently, emerged in the obesity literature. Several researchers became concerned that individuals with overweight or obesity may not perceive their weight as such and recommended correcting these perceptions to motivate weight loss [2–5]. Indeed, about one-third of youths report WPs that differ from their BMI classification [6–8]. However, consistent research indicates that such strategies are likely to have unintended and adverse consequences.
While overweight perceptions predict weight loss intentions [3, 7–11], several longitudinal studies have found perceptions of overweight to predict greater weight gain over time than “about right” perceptions [12–16]. Moreover, overweight perceptions are associated with less healthy diets, more sedentary behaviour, shorter sleep, less physical activity (PA), and an elevated risk of disordered weight-control behaviours (e.g., purging and fasting), in comparison to perceptions of being at “about the right weight” [4, 7–9, 15, 17–21]. Likewise, although they have received relatively limited research attention, perceptions of being underweight are shown to predict lower engagement in PA and less healthy diets (e.g., greater consumption of fast food, unhealthy snacks, sugar sweetened beverages, and energy drinks), than WPs of “about right.” [8, 21, 22].
Overweight and underweight perceptions also appear detrimental for youth mental health [23]. Prospective studies have found individuals who report being at “about the right weight” experience fewer depressive symptoms over time than their peers with overweight perceptions [24–26]. In fact, WPs partially account for many of the adverse psychosocial outcomes that have been associated with obesity (e.g., depression, suicidal ideation, and lower health-related quality of life) [3, 24, 27–32]. Similarly, underweight perceptions are associated with higher levels of depressive symptoms and social anxiety in males [23, 33, 34] and suicidality [27, 35] and lower health-related quality of life [29] in all youth.
Overall, perceptions of being at “about the right weight” appear advantageous for youth physical and mental health, regardless of BMI classification, whereas perceptions at either extreme (overweight or underweight) have negative effects on health behaviours and mental health. Instead of considering weight misperceptions as problematic, some researchers have proposed that underestimations of weight status may offer resiliency among individuals with overweight or obesity [26]. Promoting “about right” WPs and preventing change to overweight or underweight perceptions may offer an effective public health strategy for supporting youth health over time. However, the primary focus of WP research has been on correlates or outcomes of misperceptions, with limited prospective evidence on factors that shape perceptions of weight status. Existing literature largely consists of cross-sectional designs or differences by sociodemographic variables, but many associated variables may represent consequences of WPs rather than (or in addition to) predictors (e.g., differences in PA engagement) or are not modifiable to inform interventions (e.g., gender, age, ethnicity, and family/peer BMI). The purpose of the current study is to explore predictors of changes in WPs among youths. In particular, we examined whether bullying victimization, screen media use, and PA participation influenced the likelihood of youth changing from perceptions of being at “about the right weight” to reporting overweight or underweight perceptions over one year. We also examined whether these relationships differed by sex.
1.1. Screen Media Use
The Tripartite Influence Model posits exposure to three primary sources of influence, parents, peers, and media, contributing to the development of body image, mediated by appearance comparisons and internalization of societal appearance standards [36, 37]. Extensive body image research has examined the impact of media exposure, but the majority is exclusive to body dissatisfaction (i.e., the extent to which one experiences displeasure with his or her body) [38], rather than WPs and has been conducted in females only. The influence of screen media use likely varies by type and gender. Past body image literature largely centers around passive media viewing, such as print and television, where the concern was comparisons to celebrities and thin-idealized messages. In newer interactive forms, youth actively give and receive appearance-related feedback with peers. Extant research suggests internet use and social media have similar [39] or worse [40] effects on body satisfaction than traditional forms. Body comparisons to peers, and particularly with close friends, appear to have an equal to stronger influence on body ideals and satisfaction relative to that of celebrities [41–43]. On the other hand, social media may expose youth to the more varied body sizes of their peers, potentially having a normative effect on WPs. In support, experimental evidence shows that individuals view their bodies as smaller after exposure to normal or overweight images, relative to exposure to underweight images [44]. Female body image has generally been considered more responsive to social contexts and mass media [45–47], but this may reflect the focus on overweight concerns, without consideration of muscularity and underweight perceptions that are more common in males [6–8, 48].
1.2. Bullying Victimization
Screen use may influence WP not only through social comparisons but by exposure to cyber-bullying. Weight- or appearance-based teasing online or offline is a common experience during adolescence [49–51], reported by at least one-fifth of youth [49], and prospectively related to body dissatisfaction and monitoring in both boys and girls [52–54]. Compared to their peers, youth who have overweight and obesity are more likely to be targets of bullying and experience more frequent victimization [55–57]. Qualitative evidence suggests that underweight peers are also subject to teasing due to their underdeveloped physique, and boys report concerns about being targeted by stronger peers [58]. Some evidence suggests that links between weight status and victimization are attenuated when body satisfaction or perceived weight is accounted for [59–62]. Several cross-sectional studies demonstrate an association between bullying victimization and WPs of either overweight or underweight [61, 63–65]. Different forms of bullying engagement may lead youths to perceive their weight as overweight or underweight, with relationships potentially varied by gender, but no longitudinal studies have tested bullying as a predictor of WP changes.
1.3. PA Engagement
Body-related teasing and peer body surveillance often occur in the context of sports and other physical activities, where appearance tends to be more exposed [58, 66, 67]. Many cross-sectional studies have linked PA engagement to WPs [68, 69], which has typically been interpreted as WPs influencing PA engagement. Indeed, prospective evidence demonstrates underweight and overweight perceptions to predict lower PA than “about right” perceptions [8]; however, scant research has tested the reverse. A recent study found that adolescents who regularly practiced sports outside of school hours had more accurate body size perception than their peers not engaged in any extracurricular PA [70]. Concerns of adverse effects have typically centered around more weight-sensitive and aesthetically focused sports where thinness is considered advantageous, yet activities emphasizing muscularity also present risks for body image disturbance [71, 72]. In males, qualitative research suggests that sport and physical education settings provide important forums to compare their body to peers and discuss muscularity [58, 67, 73]. Adolescent boys report that an athletic, slim, strong, and muscular body aesthetic is necessary for sports participation, with those not conforming to this ideal subjected to teasing [67, 73]. Therefore, PA participation has the potential to be protective for WP or promote underweight or overweight perceptions, depending on context and gender.
2. Methods
2.1. Design
The COMPASS (Cannabis use, Obesity, Mental health, PA, Alcohol use, Smoking, Sedentary Behaviour) study is an ongoing (2012–2021) prospective study designed to collect hierarchical longitudinal data once annually from students in grades 9 through 12 and the secondary schools they attend [74]. School boards and schools were purposefully selected based on whether they permitted active-information passive-consent parental permission protocols [74], which are critical for collecting robust data among youth [75]. All grade 9 through 12 students attending participating schools were eligible to participate and could decline at any time. A full description of COMPASS and its methods are available in print [74] or online (http://www.compass.uwaterloo.ca). All procedures were approved by the University of Waterloo and Brock University Office of Research Ethics and appropriate school board committees.
2.2. Participants
The current study used linked student-level data from Year 3 (Y3: 2014–2015) and Year 4 (Y4: 2015–2016) of the study. In Y3, data were collected from 42,355 youth (79.3% participation rate) in 87 secondary schools in Ontario (n = 39,013 at 78 schools) and Alberta (n = 3342 at 9 schools). In Y4, data were collected from 40,436 students in 81 schools (9 Alberta schools; 72 Ontario schools (7 withdrew due to labour issues/illness, 1 new school)) [76]. Missing respondents resulted primarily from scheduled free/study periods or absenteeism during data collection.
To explore longitudinal changes among respondents, Y3 and Y4 student-level data were linked within schools (80 schools participated in both Y3 and Y4). The process of linking student data across waves is described in more detail by Qian and colleagues [77]. Due to the rolling sample design [74], it was not possible to link students who were in grade 12 at first participation and graduated that year or the grade 9 students that were newly admitted to participating schools in Y4. The other main reasons for nonlinkage included students transferring schools or dropping out, students not providing data for grade or sex (232), students on scheduled free/study periods or absent during data collection, or inaccurate data provided in the linkage measures. A total of 17,880 students were successfully linked for the two years of data collection. The final sample consisted of 17,475 youth, after removing students missing WP data for Y3 (237) and/or Y4 (184).
2.3. Data Collection Tool
Student-level data were collected using the COMPASS student questionnaire (Cq), a paper-based survey designed to collect student-reported data on multiple health behaviours, correlates, and demographic variables from full school samples during one classroom period. Cq items were based on the national standards or current national public health guidelines as described elsewhere [74]. The cover page contains measures to create a unique self-generated code for each respondent in a school to ensure the anonymity of participants, while still allowing COMPASS researchers to link each student’s anonymous identifier data over multiple years [78].
2.4. Measures
2.4.1. WP
Consistent with previous studies [8], WP was assessed by asking “how do you describe your weight?” Response options included the following: “very underweight,” “slightly underweight,” “about the right weight,” “slightly overweight,” and “very overweight.” Responses of “very underweight” and “slightly underweight” and of “very overweight” and “slightly overweight” were collapsed into “underweight” and “overweight,” respectively. Changes from perceptions of “about the right weight” to perceptions of “overweight” or to “underweight” were modelled.
2.4.2. PA Measures
The PA measures have been previously validated [79]. To assess moderate-to-vigorous PA (MVPA), respondents were asked how many minutes of hard- and moderate-intensity PA they engaged in on each of the last 7 days. Consistent with the Canadian PA guidelines for youth [80], students were classified based on whether they had performed at least 60 minutes of daily MVPA on each of the last 7 days. Similarly, students were categorized based on whether they met the three times weekly recommendation for resistance exercise [81], by asking “on how many days in the last 7 days did you do exercises to strengthen or tone your muscles (e.g., push-ups, sit-ups, and weight training)?” Other PA items assessed whether respondents participated in competitive sports teams against other schools (e.g., varsity sports), league or team sports outside of school, and school-organized PA at noon, before, or after school (e.g., intramurals and noncompetitive clubs).
2.4.3. Screen Use
Using previously validated measures [82], screen time was assessed by asking students the average time in hours and minutes per day that they spent: “watching/streaming TV shows or movies,” “playing video/computer games,” “talking on the phone,” “surfing the internet,” and “texting, messaging, emailing.” Responses were categorized based on whether they exceeded two hours per day on average for each type of screen use.
2.4.4. Bullying Victimization
Bullying victimization was assessed by the following question: “in the last 30 days, in what ways were you bullied by other students? (mark all that apply).” Provided response options included the following: “I have not been bullied in the last 30 days,” “physical attacks (e.g., getting beaten up, pushed, or kicked),” “verbal attacks (e.g., getting teased, threatened, or having rumours spread about you),” “cyber-attacks (e.g., being sent mean text messages or having rumours spread about you on the internet),” and “had someone steal from you or damage your things.” Responses were dichotomized according to whether they reported having experienced each form of bullying in the last 30 days.
2.4.5. Covariate and Stratifying Measures
Models were adjusted for student-reported race/ethnicity (white, nonwhite minority (black, Asian, indigenous [Métis, Inuit, First Nations], Hispanic/Latin American, others), grade (9–12), and BMI (kg/m²) category (recoded as underweight, normal weight, overweight/obesity, and missing). BMI classifications were determined based on student-reported height and weight [83] and the World Health Organization’s [78] age- and sex-adjusted cutoff points. The weight status measure has been found to be reliable, valid, and valuable for use when objective methods are not feasible [84]. Missing BMI was included as category, due to the high amount of missing data (due to missing height, weight, age, or sex). Regression models were stratified by student-reported sex (male, female).
2.5. Analyses
All analyses were performed using the statistical package SAS 9.4. Descriptive statistics were calculated by sex. Generalized Estimating Equation (GEE) models were used, with independent working correlation for multinomial outcomes [85]. The GEE model is an extension of generalized linear models to correlated data, simply modelling the mean response and treating covariance as nuisance. It produces consistent estimates for regression parameters. Models tested Y4 measures of types of PA participation, bullying victimization, and screen use as predictors of changes to Y4 WPs of “overweight” or “underweight” (reference category: Y4 “about the right weight”), adjusting for Y3 covariates (grade, ethnicity, BMI classification), in youth with a Y3 WP of “about the right weight.” Models were stratified by sex and adjusted for Y3 covariates (grade, ethnicity, BMI classification). Schools were included in the models as clusters to take account of within-school correlation.
3. Results
3.1. Descriptive Statistics
Descriptive statistics for all variables in Y3 are presented in Table 1. The majority of the sample identified as white (72.6% females; 68.3% males) and 52.3% as indicated they were female. Females were more likely than males to report experiences of verbal bullying or cyber-bullying victimization in the last 30 days (16.9% vs. 11.9% and 7.5% vs. 2.2%, respectively), while males were more likely to report physical attacks (3.2% vs. 1.2%) or having their belongings stolen or damaged (3.1% vs. 2.3%) than females. More males than females met the MVPA (58.1% vs. 47.5%) and resistance exercise guidelines (57.4% vs. 50.3%) and played competitive (41.9% vs. 37.6%) and noncompetitive school sports (48.4% vs. 39.6%), while more females played sports outside of school than males (45.3% vs. 39.9%). A higher portion of males played video/computer games in excess of 2 hours a day (49.1% vs. 10.5%); females outnumbered males in exceeding 2 hours per day for all other forms of screen use.
Female (n = 9225)
Male (n = 8655)
Chi-square
% (n)b
% (n)b
value
Grade
9
35.7 (3295)
38.0 (3286)
<0.0001
10
34.8 (3209)
34.5 (2989)
11
28.2 (2601)
25.6 (2214)
12
1.3 (120)
1.9 (166)
Ethnicity
White
72.6 (6698)
68.3 (5911)
<0.0001
Nonwhite
27.4 (2527)
31.7 (2744)
BMI classificationa
Underweight
1.3 (117)
1.6 (135)
<0.0001
Normal weight
59.5 (5492)
51.9 (4490)
Overweight/obesity
15.0 (1383)
25.6 (2217)
Missing BMI
24.2 (2233)
20.9 (1813)
WP
Underweight
11.0 (1006)
21.9 (1864)
<0.0001
“About right”
57.5 (5237)
55.9 (4767)
Overweight
31.5 (2872)
22.2 (1897)
PA
School-organized PA (e.g., intramurals)
Yes
39.6 (3631)
48.4 (4135)
<0.0001
No
60.4 (5535)
51.6 (4415)
School league/team sports (e.g., varsity)
Yes
37.6 (3445)
41.9 (3588)
<0.0001
No
62.4 (5710)
58.1 (4974)
League/team sports outside of school
Yes
47.5 (4338)
58.1 (4975)
<0.0001
No
52.5 (4797)
41.9 (3587)
Met MVPA guidelinesc
Yes
47.5 (4338)
58.1 (4975)
<0.0001
No
52.5 (4797)
41.9 (3587)
Met resistance exercise guidelinesc
Yes
50.3 (4601)
57.4 (4900)
<0.0001
No
49.7 (4552)
42.6 (3644)
Screen time
TV/movies
≤2 hours/day
43.6 (4012)
47.6 (4112)
<0.0001
>2 hours/day
56.4 (5200)
52.4 (4526)
Video/computer games
≤2 hours/day
89.5 (8248)
50.9 (4398)
<0.0001
>2 hours/day
10.5 (964)
49.1 (4240)
Surfing the Internet
≤2 hours/day
48.1 (4432)
61.7 (5327)
<0.0001
>2 hours/day
51.9 (4780)
38.3 (3311)
Talking on the telephone
≤2 hours/day
89.1 (8204)
94.0 (8117)
<0.0001
>2 hours/day
10.9 (1008)
6.0 (521)
Texting/messaging/emailing
≤2 hours/day
48.1 (4429)
67.6 (5842)
<0.0001
>2 hours/day
51.9 (4780)
32.4 (2796)
Bullying victimization in the last 30 days
Physical
Yes
1.2 (107)
3.2 (273)
<0.0001
No
98.8 (9118)
96.8 (8382)
Verbal
Yes
16.9 (1557)
11.9 (1031)
<0.0001
No
83.1 (7668)
88.1 (7624)
Cyber
Yes
7.5 (692)
2.2 (191)
<0.0001
No
92.5 (8533)
97.8 (8464)
Stolen/damaged belongings
Yes
2.3 (214)
3.1 (265)
<0.0001
No
97.7 (9011)
96.9 (8390)
aBMI classification based on self-reported height and weight and age- and sex-adjusted cutoffs. bNumbers may not add to total due to rounding and/or missing data. cBased on PA guidelines for youths doing at least 60 min/day of MVPA and strengthening exercises at least 3 days/week.