Bullying Experiences, Body Esteem, Body Dissatisfaction, and the Moderating Role of Weight Status Among Adolescents

Abstract

Introduction:

The present study examined the relationships between bullying victimization experiences, body esteem (BE), and body dissatisfaction (BD) by obesity status.

Methods:

Using a cross-sectional design, adolescents (10-16y) from the U.S. completed surveys. Weight and height were objectively assessed. Bullying was assessed using the Personal Experiences Checklist, with a total score that included three subscales (cyber, verbal-relational, and physical). BE was assessed using total score on the BE Scale, and BD was assessed using the Body Image Assessment for Pre-Adolescents, operationalized as the discrepancy between current and ideal body figure. Linear regression models, controlling for sociodemographics, were used to examine obesity status (with vs. without obesity) and sex as moderators of the relationships between bullying, BE, and BD.

Results:

On average, adolescents (n=328) were 12.6±1.9 y, 45.4% male, 58.2% White, 35.7% Black, and 36.0% had obesity. Obesity status moderated the relationship between bullying experiences and BE, such that these relationships were only significant (B=−0.38, p<.05) among adolescents with obesity compared to adolescents without obesity. In both groups, greater verbal-relational bullying experiences were associated with lower BE, while greater physical bullying experiences were related to higher BE only among adolescents with obesity. In adolescent girls without obesity, greater cyberbullying experiences were associated with lower BD.

Conclusions:

The findings suggest different associations between sub-types of bullying victimization experiences and BE and BD, highlighting the importance of examining these associations separately by obesity status.

Introduction

Adolescents experience rapidly developing perceptions, cognitions, and feelings toward one’s changing body (McCabe & Ricciardelli, 2004), contributing to sense of self (Harter, 2003; Rosenblum & Lewis, 1999). Further, peer interactions change during this time (Markey, 2010), contributing to an adolescent’s body image (Jones & Crawford, 2006). Body image is a multifaceted construct that includes perceptions of one’s own body size and feelings, cognitions, and behaviors in relation to one’s body (Neves et al., 2017). Given that body image disturbances are associated with poor mental health and negative health outcomes such as eating psychopathology and weight gain (Duarte et al., 2017; Pallan et al., 2011), it is important to understand factors associated with poor body image, particularly during adolescence when body image is developing.

The diathesis-stress model offers a framework for delineating the influences on body image. In this model, negative life experiences such as bullying victimization interact with an adolescent’s personal vulnerabilities (i.e., obesity status) to influence body image (Flores et al., 2019; Swearer & Hymel, 2015). Pediatric obesity, defined as a body mass index (BMI) for age and sex above the 95^th percentile (Kuczmarski et al., 2000), continues to be a public health concern, affecting approximately one in five youth 2 to 19 years of age (Skinner et al., 2018). Adolescents with obesity often have worse body image disturbances compared to adolescents with normal or overweight status (Danielsen et al., 2012; Flores et al., 2019; Goldfield et al., 2010; Griffiths et al., 2010; Shin & Shin, 2008). Further, adolescents with obesity are at risk for negative physical and mental health outcomes (Rankin et al., 2016) including bullying victimization (Lumeng et al., 2010).

Body image may relate differentially to weight, depending on the aspect of body image being measured or examined (Tyler et al., 2008). Body esteem (BE) reflects the attitudinal dimension of body image, i.e., self-reported attitudes, beliefs, and body appreciation (Mendelson et al., 2001). Body dissatisfaction (BD), a perceptual dimension of body image, refers to the extent to which one’s perceived body shape and size are different from one’s ideal body (Veron-Guidry & Williamson, 1996). While both BE and BD are associated with poor mental and physical health outcomes (Goldfield et al., 2010; Neves et al., 2017), each may require different intervention approaches. For example, orientation to healthy body sizes might be more of a focus to improve BD whereas body appreciation may be the greater emphasis in BE-targeted interventions (Cantone et al., 2015). Additionally, these constructs can change in different ways; for example, weight loss might lead to increases in BE but not necessarily decreases in BD, as the perception of an ideal body shifts toward a thinner standard, causing BD to persist (Russell & Cox, 2003).

Unfortunately, being a victim of bullying is a common experience in adolescence, with approximately one in five U.S. 12-to-18-year-olds reporting being bullied at school (Musu et al., 2019). Bullying is persistent, aggressive behavior characterized by an imbalance of power and directed with the intention of causing harm (Gladden et al., 2014) and can be expressed in many forms, such as physical (e.g., pushing), relational (i.e., efforts to harm the reputation or relationships of the targeted youth), or bullying that occurs through the use of technology (i.e., cyberbullying). While sub-types of bullying victimization often co-occur (Wang et al., 2010), they also may have differential effects on psychological health outcomes (Brunstein Klomek et al., 2019). For example, having a single physical, but not verbal or relational, bullying victimization experience was associated with suicidal ideations and attempts, while chronic victimization was differentially predictive of ideations and attempts depending on the type of bullying (Brunstein Klomek et al., 2019). There are also significant short- and long-term deleterious effects of bullying victimization on adolescents (Lian et al., 2018; Ranta et al., 2009). Related to weight, bullying among adolescents is prospectively related to increased BMI (Rosenthal et al., 2015; Schvey et al., 2019) and the probability of overweight five years later (Haines et al., 2007).

Previous cross-sectional studies have demonstrated that the relationship between weight status and exposure to bullying can be partially explained by changes in body image or BE (Brixval et al., 2011; Fox & Farrow, 2009). Both studies proposed that adolescents with obesity are at greater risk of exposure to bullying and that decreased body image (Brixval et al., 2011) and increased BD (Fox & Farrow, 2009) mediate the relationship between weight status and exposure to teasing. Similarly, prospective studies have suggested that BD may increase risk for future peer victimization more than weight status (Sutter et al., 2015). Longitudinal research also suggests that weight-related teasing in childhood is related to BD in female adults (Jackson et al., 2000).

Though exposure to bullying may differ by weight status, investigations into differences in body image disturbances by specific types of bullying between adolescents with and without obesity is limited. For example, cyberbullying experiences may be associated with greater BD among adolescents (Kenny et al., 2018), but it is unclear if this relationship differs by weight status. In testing the moderating role of healthy vs. overweight/obesity on the relationship between bullying types (i.e., verbal, physical, social, cyber), previous work has not separated youth with obesity from youth with overweight (Fox & Farrow, 2009), warranting examination of obesity status separately from overweight status as a predictor. Delineating these relationships can provide more specific guidance for interventions.

Finally, associations between body image and weight among adolescents may differ by gender (Calzo et al., 2012; Fernández-Bustos et al., 2020). First, previous work suggests that adolescent girls, compared to adolescent boys, may have more negative body image (Kantanista et al., 2015), lower BE (Mak et al., 2013), and greater BD (Fernandez-Bustos et al., 2019). Further, gender has also been shown to be an important moderator of associations between BMI and BD among adolescents, such that adolescent females may have increased BD as BMI increases (i.e., a linear relationship between BMI and BD) while adolescent males may have greater BD at BMIs classified as underweight or overweight (i.e., a curvilinear relationship between BMI and BD) (Austin et al., 2009). Associations between different forms of peer victimization experiences (e.g., social exclusion, appearance teasing) and different aspects of BE (e.g., weight-related evaluations, beliefs about how others perceive one’s appearance) have also been shown to differ by gender (e.g., bullied girls were shown to have more negative beliefs about how others perceive their appearance than bullied boys) (Lunde et al., 2006). Therefore, it is important to consider the role of gender when examining associations between body image, bullying experiences, and weight.

Present Study

Both bullying victimization and weight status may be related to BE and BD (Brixval et al., 2011; Fox & Farrow, 2009), and youth with obesity may be at greater risk of bullying victimization (Lumeng et al., 2010); however, differences by weight status and exploration of specific types of bullying are needed. Further, few studies have used a diathesis-stress model of bullying to better understand how bullying experiences might serve as stressful events that influence the relationships between weight status and body image cognitions. Thus, the present study aimed to delineate the relationships between bullying experiences, BE, and BD and determine whether the relationships differ based on obesity status. The study proposed two hypotheses:

Hypothesis 1: It was hypothesized that obesity status will moderate the relationship between bullying victimization experiences and BE and BD. More specifically, bullying experiences will be inversely related to BE, and positively related to BD, based on prior literature (Fox & Farrow, 2009). Further, the relationships between bullying experiences and BE and between bullying experiences and BD will be moderated by weight status, such that these relationships will be strongest among adolescents with obesity vs. without obesity.

Hypothesis 1a. It was also hypothesized that the relationships will be stronger among females compared to males, based on prior literature (Calzo et al., 2012; Fernández-Bustos et al., 2020). Hypothesis 2: Relationships between different types of bullying victimization (i.e., physical, verbal/relational, or cyberbullying) and BE, and bullying and BD, will be different for adolescents without obesity vs. with obesity.

Method

Participants

Adolescents between 10-16 years of age were recruited between 2016-2018 from a metropolitan city in the southeastern U.S. for the TIGER Kids study, a prospective observational cohort study examining the relationships between physical activity, the socio-emotional context, and obesity. The current study used baseline measurements. Using a convenience sampling strategy, participants were recruited via website advertisements, email listservs, and flyers in the greater community, including school events and pediatric clinics. Exclusion criteria included being on a restrictive diet due to illness, weight greater than 500 pounds, pregnancy, or significant physical/mental disabilities that impede walking.

Procedure

Parents provided written consent for adolescents, and children provided assent. Parents reported the adolescent's birth date, sex, race/ethnicity, and household income. Adolescents completed questionnaires including to collect information on whether they were in school or out of school (i.e., summer vacation) and the surveys described below using REDCap (Research Electronic Data Capture), a secure, web-based application designed to support data capture for research studies (Harris et al., 2009). Research staff measured adolescent weight (to the nearest 0.1 kilogram) and height (to the nearest 0.1 centimeter) using an electronic scale and wall-mounted stadiometer in a gown and barefoot, and adolescents were asked to arrive fasted. Measurements were taken twice with the analyzed value being the average of the measurements. A third measurement was taken if the two measurements differed by more than 0.5 units. The Institutional Review Board approved the study (#2016-028).

Measures

Bullying.

Bullying victimization was assessed using the Personal Experiences Checklist (Hunt et al., 2012), a 32-item questionnaire including four subscales of bullying (verbal-relational, physical, cyber, and cultural). Participants reported how often, in the past month, they experienced (i.e., were a victim of) different types of bullying at school (e.g., verbal-relational: “Other kids say mean things about me behind my back;” physical: “Other kids shove me;” cyber: “Other kids say nasty things about me on an instant messenger or chat room”), on a five-point scale from Never (1) to Everyday (5). The measure was reliable and valid among Australian adolescents 8 to 16 years of age (Hunt et al., 2012). Previous work has demonstrated support for the use of subscales as well as a total bullying score (Hunt et al., 2012). Given that culture-related items may be specific to the population it has been validated in (e.g., Australian adolescents), the cultural bullying subscale was not included in the present study. Therefore, total scores were the sum of 28 items (verbal-relational, physical, and cyber bullying subscales), with greater scores reflecting more bullying experiences; possible scores range from 28 to 140.

Body Esteem.

BE was measured using total score on the BE Scale (Mendelson et al., 2001). The scale was found to be reliable and valid among adolescents (Cecil & Stanley, 1997). Participants indicated how often they agreed with 23 statements related to their body (e.g., “I like what I see when I look in the mirror”), on a scale from Never (1) to Always (5). Potential scores range from 23 to 115, with greater scores reflecting higher BE. Along with total BE scores, the BE Scale includes three subscales: (a) general feelings about appearance (appearance-related, 10 items), (b) weight satisfaction (weight-related, 8 items), and (c) how an individual thinks others evaluate his or her body (i.e., attribution-related, 5 items).

Body Dissatisfaction.

BD was assessed using the Body Image Assessment for Pre-Adolescents (BIA; Veron-Guidry & Williamson, 1996). Participants were asked to indicate, using nine body silhouettes specific to adolescent’s sex, which of the images is most reflective of their current figure and which image corresponds to their ideal figure. BD was operationalized as the discrepancy between current body figure and ideal body figure, with scores closest to zero indicating low BD, negative scores reflecting a thinner current image than ideal image, and positive scores reflecting a thinner ideal image than current image. The BIA has been validated among adolescents (Truby & Paxton, 2002), and it is recommended to analyze scores separately for boys and girls due to the sex-specific silhouettes (Veron-Guidry & Williamson, 1996).

Puberty Status.

Tanner stage of development was used to assess puberty status. Adolescents self-reported their stage of sexual maturation, based on a series of standardized, validated drawings depicting progressive stages of pubertal development from 1 (no development) to 5 (complete development) for female breasts or male genitalia and from 1 (no development) to 6 (complete development) for pubic hair development (Tanner, 1986). Puberty status was categorized based on published recommendations and developed for a cardiometabolic risk score into pre- (stage 1), peri- (stages 2 – 4), and post-puberty (stages 5 – 6) (Eisenmann, 2008; Marshall & Tanner, 1970).

Data Analysis

Participants who provided complete measures for the demographics, anthropometrics, bullying experiences, BE, and BD were included in analyses. Adolescents were categorized into weight status groups based on their age- and sex-specific BMI percentile (Kuczmarski et al., 2000), including underweight (<5^th percentile for age- and sex-specific BMI), normal (≥5th and <85^th percentile), overweight (≥85th and <95^th percentile), class I obesity (≥95^th and <120% of the 95^th percentile), and severe obesity (i.e., class II and III, ≥120% of the 95^th percentile (Flegal et al., 2009). One-way ANOVA was used to examine the differences in subscales and total scores on the PECK, BE, and BIA across weight status groups (normal, overweight, obesity, and severe obesity as few adolescents had underweight), using a Brown-Forsythe test for homogeneity of variance, with Tukey HSD post hoc tests performed if significant differences (p<0.05) were found. If homogeneity of variance was violated (p<0.05), Welch’s ANOVA was used for group means. Spearman rank correlation coefficients were conducted between subscales and total scores of the PECK, BE, BIA, and BMI percentile. Covariates included age, sex, race (White, Black, and “other”), parents’ marital status, household size, in/out of school status, and puberty status (pre, peri, and post). For analysis, total bullying scores were centered to the median (score of 31) due to skewness. Significance for main effects in linear regression was set at p<0.05. All analyses were performed in SAS version 9.4 (Cary, N.C.)

For the first hypothesis, general linear regression models were used to assess the five separate models of the relationship between total bullying (predictor), obesity weight status (0 = BMI percentile < 95^th, 1= BMI percentile ≥ 95^th, moderator; Kuczmarski et al., 2000), and outcomes (total BE, BE subscales [appearance, attribution-related, and weight], and BD) including covariates. Dichotomization of obesity status was chosen to directly test the moderating role of obesity status, given the increased risk of poor body image and bullying victimization among adolescents with obesity. All models examining BD as the outcome were examined in boys and girls separately. Further, sex and obesity status were tested as moderators between total bullying, total BE, and BE subscales. The potential clustering effect of siblings within a shared household was explored using mixed model analysis. Model outcomes were identical between mixed models when controlling for the data nested by household and non-clustered linear regression; therefore, linear model results are presented. Q-Q plots and residual plots were examined to test the assumptions of linear regression, indicating that assumptions were not violated.

For the second hypothesis, five separate linear regression models were used to assess the relationship among bullying subscales (cyber, physical, relational), with total BE and BD stratified by sex and obesity status (i.e. males with/without obesity, females with/without obesity). Bullying subscales were centered on the median for these analyses. Backward selection was conducted to obtain the most parsimonious model, starting with all three bullying subscales and covariates used in the first hypothesis, concluding with all variables in the model with a p-value <0.05.

Results

Of the 342 participants, 328 provided complete measures excluding participants with incomplete PECK questionnaires (n=2), incomplete weight measurements (n=2), underweight (<5^th percentile; n=9), and missing puberty status (n=1).¹ On average, adolescents were 12.6 (SD=1.9) years of age, 58.2% were White, 35.7% were African American, and 6.1% reported other race. Most participants (92.4%) were non-Hispanic, 36.0% had obesity, and most (52.4%) were peri-puberty (Table 1).

Table 1.

Descriptive Characteristics of Adolescents (n=328).

	Mean±SD	n (%)
Age (years)	12.6±1.9
Sex, Male		149 (45.4)
Race,
White		191 (58.2)
African-American		117 (35.7)
Other		20 (6.1)
Ethnicity,
Hispanic		25 (7.6)
Non-Hispanic		303 (92.4)
Annual household income
< $29,000		36 (11.0)
$30,000 – 70,000		83 (25.3)
$71,000 – 140,000		115 (35.1)
>$140,000		74 (22.6)
Did not Answer		20 (6.1)
In school (vs. on school holiday)		199 (60.7)
Puberty Stage
Pre-puberty		34 (10.4)
Peri-puberty		172 (52.4)
Post-puberty		122 (37.2)
Bullying, Body Esteem, and Body Image
Bullying Experiences
Cyber Bullying	8.8±2.5
Physical Bullying	10.8±4.1
Verbal-relational Bullying	15.5±6.8
Total Experiences	35.1±11.8
Body Esteem
Appearance-related	27.3±9.4
Attribution-related	9.9±4.7
Weight-related	19.5±6.5
Total	56.7±13.0
Body Image Assessment
Current Figure	4.5±1.7
Ideal Figure	3.4±1.1
Body Dissatisfaction	1.1±1.6
Weight Status
BMI Percentile	74.1±27.7
Weight Classificationa
Normal		158 (48.2)
Overweight		52 (15.9)
Obesity		54 (16.5)
Severe Obesity		64 (19.5)

BMI = Body Mass Index

^aWeight classes are mutually exclusive where normal is BMI percentile >5^th and <85^th for child age and sex, overweight is ≥85^th and <95th percentile, obesity is ≥95^th percentile and <120% of the 95^th percentile, and severe obesity is ≥120% of the 95^th percentile

As shown in Table 2, when comparing differences between weight groups (with and without obesity) total bullying and other subscales of bullying, the appearance-related, attribution-related, weight-related, and total BE were all significantly different across weight groups (ps<0.001; omnibus F statistics and pairwise differences are presented in Table 2). Welch’s ANOVA was used for differences in group means by attribution-related, weight-related, and total BE, along with current BIA, as homogeneity of variances was violated. All BE measures (appearance-related, attribution-related, weight-related, and total) were lowest among adolescents with severe obesity and highest among those with normal weight. Pairwise comparisons showed that weight groups differed significantly on all BE and BD measures (ps<0.001). Adolescents with severe obesity differed from those with normal weight and overweight in appearance-related (difference scores ranged from 6.25 - 7.51), attribution-related (difference scores ranged from 2.29 – 2.49), and total BE (difference scores ranged from 12.40 – 18.79; p<0.05 for all). Adolescents with severe obesity differed from all other groups in weight-related BE and BD (p<0.05). Notably, there were large effect sizes between adolescents with severe obesity from normal weight adolescents in regards to total BE (cohen’s d=2.96, mean difference=18.79, 95% CI = 12.57 to 25.00) and BD (cohen’s d=4.86, mean difference= −2.42, 95% CI = −2.91 to −1.93).

Table 2.

Descriptive statistics and differences across weight status (n=328)^{^}.

	Normal (n=158)	Overweight (n=52)	Obesity (n=54)	Severe Obesity (n=64)	F	p-value
Bullying Experiences
Cyber Bullying	8.9±3.0	8.7±1.4	8.2±0.7	9.0±2.7	0.48	0.69
Physical Bullying	10.9±4.4	10.8±4.8	10.2±2.2	11.1±4.2	0.51	0.67
Relational Bullying	14.9±5.7	16.4±9.1	14.3±5.7	17.3±7.8	2.38	0.07
Total Experiences	34.7±11.9	35.9±13.9	32.8±7.3	37.5±12.3	1.74	0.16
Body Esteem
Appearance-related	30.2±8.3a	27.6±8.1a,b	23.9±10.9b,c	22.7±9.1c	13.34	<.001*
Attribution-related	10.7±4.2a	10.5±5.4a	8.9±4.9a,b	8.3±4.9b	5.61	0.001*
Weight-related	22.9±4.6a	19.3±5.5b	16.3±6.9c	14.1±6.2c	44.18	<.001*
Total	63.8±13.8a	57.4±16.3a	49.1±20.2b	45.0±18.0b	23.43	<.001*
Body Image Assessment
Current Figure	3.4±1.0a	4.8±1.2b	5.1±1.4b,c	6.6±1.4d	102.99	<.001*
Ideal Figure	3.2±0.9a	3.2±1.0a	3.3±1.1a	4.0±1.3b	10.05	<.001*
Body Dissatisfaction	0.2±1.2a	1.6±1.2b	1.8±1.4b	2.6±1.5c	65.03	<.001*

^{^}Assessed using One-Way ANOVA or Welch’s ANOVA

^*p<0.05.

Means with identical subscripts (_a,b,c,d) are not significantly different at p<0.05

When examining correlations amongst bullying scales, BE, BD, and covariates, total bullying experiences and bullying subscales (cyber, physical, and relational) were significantly related to each other (ps<0.001). Similarly, total BE and BE subscales were significantly related to each other (ps<0.001). Total BE and appearance-related BEwere negatively correlated with total bullying and subscales (ps<0.05), and attribution-related BE and weight-related BE were related to verbal-relational bullying (ps<0.05). Attribution-related BE was negatively related to total bullying, and weight-related BE was negatively related to cyberbullying (ps<0.05). BD was negatively associated with total BE and all BE subscales (ps<0.001). Percent of the 95^th BMI percentile was positively related to BD and negatively related to total BE and all BE subscales (ps<0.001).

Hypothesis 1: Moderation Model of Bullying, Body Esteem, and Body Dissatisfaction by Obesity Status and Sex

Obesity status was consistently related to total BE, all subscales of BE, and BD (ps<.001), while in-school status, household quantity, puberty status, and sex were all not associated with any of the outcomes in models (ps>.05; Table 4). In models stratified by obesity status, including participants with obesity, total bullying was negatively associated with total BE in adolescents with obesity (unstandardized B=−0.38, 95% CI = −0.69, −0.06, p=.02) but was not associated with total BE in adolescents without obesity (B=−0.13, 95% CI = −0.29, 0.04, p=0.14; Figure 1).

Table 4.

Unstandardized beta coefficients for the associations of bullying experiences with body esteem and body dissatisfaction (n=328)^{^}.

	Appearance-Related Esteem			Attribution-Related Esteem			Weight-Related Esteem
	Beta	95% CI	p-value	Beta	95% CI	p-value	Beta	95% CI	p-value
Total Bullying	−0.15	−0.24, −0.05	0.002*	0.01	−0.04, 0.06	0.756	−0.01	−0.07, 0.06	0.821
OB	−6.32	−8.42, −4.22	<0.001*	−2.36	−3.48, −1.25,	<0.001*	−6.75	−8.14, −5.36	<0.001*
OB*Total Bullying	−0.13	−0.31, 0.04	0.133	−0.07	−0.16, 0.03	0.169	−0.08	−0.19, 0.04	0.185
Male	−0.44	−2.36, 1.49	0.658	0.21	−0.82, 1.23	0.695	−0.74	−2.01, 0.53	0.256
Age	−0.97	−1.57, −0.38	0.002*	0.12	−0.19, 0.44	0.447	−0.50	−0.90, −0.11	0.013*
Race
White	Reference
Black	2.59	0.48, 4.71	0.017*	1.31	0.18, 2.43	0.024*	1.12	−0.29, 2.52	0.120
Other	0.51	−3.52, 4.54	0.804	−1.90	−4.06, 0.24	0.083	1.50	−1.17, 4.17	0.271
Parent Marital Status
Married	Reference
Divorced or Separated	−0.31	−2.64, 2.02	0.796	−0.52	−1.76, 0.72	0.413	−0.21	−1.75, 1.33	0.792
Other	0.30	−2.45, 3.05	0.831	1.61	0.15, 3.07	0.032*	0.62	−1.20, 2.44	0.503
In-school Status	1.10	−0.83, 3.03	0.265	0.64	−0.39, 1.67	0.224	1.24	−0.04, 2.51	0.059
Puberty Status
Pre-puberty	0.95	−2.93, 4.82	0.632	−1.33	−3.39, 0.73	0.205	−1.65	−4.22, 0.91	0.208
Peri-puberty	−0.40	−2.69, 1.90	0.736	−0.65	−1.87, 0.56	0.294	−0.29	−1.80, 1.23	0.712
Post-puberty	Reference
Household Size	0.11	−0.58, 0.80	0.756	−0.28	−0.65, 0.09	0.136	0.10	−0.36, 0.55	0.684
R squared	0.23			0.14			0.30
	Total Body Esteem			Body Dissatisfaction (Boys)			Body Dissatisfaction (Girls)
	Beta	95% CI	p-value	Beta	95% CI	p-value	Beta	95% CI	p-value
Total Bullying	−0.15	−0.33, 0.03	0.104	0.00	−0.03, 0.02	0.807	−0.01	−0.03, 0.01	0.337
OB	−15.43	−19.39, −11.47	<0.001*	1.51	0.91, 2.10	<0.001*	1.58	1.16, 1.99	<0.001*
OB*Total Bullying	−0.28	−0.61, 0.05	0.099	0.01	−0.04, 0.07	0.643	0.03	0.01, 0.07	0.038*
Male	−0.97	−4.60, 2.66	0.601
Age	−1.35	−2.48, −0.23	0.019*	−0.06	−0.25, 0.13	0.554	−0.05	−0.16, 0.06	0.410
Race
White	Reference
Black	5.02	1.02, 9.01	0.014*	0.01	−0.60, 0.61	0.987	0.02	−0.41, 0.44	0.938
Other	0.11	−7.50, 7.72	0.977	−0.34	−1.70, 1.02	0.627	0.22	−0.50, 0.94	0.550
Parent Marital Status
Married	Reference
Divorced or Separated	−1.03	–5.43, 3.36	0.646	−0.13	−0.77, 0.51	0.684	−0.13	−0.61, 0.34	0.578
Other	2.53	–2.66, 7.72	0.339	0.59	−0.15, 1.32	0.122	0.02	−0.55, 0.58	0.957
In-school Status	2.97	–0.67, 6.62	0.111	−0.16	−0.68, 0.35	0.537	−0.02	−0.42, 0.38	0.920
Puberty Status
Pre-puberty	−2.04	–9.35, 5.27	0.585	−0.56	−1.65, 0.53	0.316	−0.04	−0.92, 0.83	0.921
Peri-puberty	−1.34	–5.66, 2.99	0.546	−0.44	−1.22, 0.33	0.262	−0.25	−0.66, 0.16	0.237
Post-puberty	Reference
Household Size	−0.08	–1.38, 1.23	0.910	0.10	−0.10, 0.33	0.324	−0.09	−0.23, 0.04	0.182
R squared	0.25			0.24			0.37

OB = Obesity or severe obesity weight status = reference group; CI = Confidence Interval.

^*indicates p<0.05

^{^}Assessed using linear regression with the following covariates: age, sex, race, puberty status, household quantity, in-school status, and parental marital status

Figure 1.

Association between Total Body Esteem and Total Bullying Experiences by Obesity Status.

Solid line represents adolescents with healthy or overweight status, and dashed line represents adolescents with obesity or severe obesity; Assessed using linear regression with adjustment for age, sex, race, in-school status, puberty status, household size, and parent’s marital status.

*Beta coefficient is significantly different from zero at p<.05

Obesity status did not moderate the relation between bullying and BD among boys. The interaction between girls’ obesity status and bullying was significant for BD (model R²=0.37, B=0.03, 95% CI= 0.01 to 0.07, p=0.038); however, in models stratified by obesity status, bullying was not significantly associated with BD in girls with or without obesity (ps>.05). Sex was not a significant moderator of the interactions between total bullying and obesity status associated with total BE, appearance-related BE, or weight-related BE (ps>.10). The three-way interaction between sex, obesity status, and bullying was significant for attribution-related BE (model R²=0.34) such that there was an inverse relationship between bullying experiences and attribution-related BE only for boys with obesity (B=−0.14, 95% CI= −0.27 to −0.01, p<.05).

Hypothesis 2: Bullying subscales with Body Esteem and Body Dissatisfaction by Obesity Status

The relationships between different types of bullying and BE and BD were different for adolescents without vs. with obesity, supporting Hypothesis 2 (see Table 5). After backwards selection, greater verbal-relational bullying experiences were related to lower total BE (B=−0.32, 95% CI=−0.61 to −0.03, p=0.033), in adolescents without obesity but only accounted for 6% of the variance in total BE. For adolescents with obesity, greater verbal-relational bullying was associated with lower total BE (B=−0.99, 95% CI= −1.53 to −0.46, p<0.001), and more physical bullying experiences were related to greater total BE (B=1.23, 95% CI= 0.11 to 2.35, p=0.031) with the final model accounting for 23% of the variance in total BE, including race (greater BE among Black adolescents compared to White adolescents, p=0.004) and parental marital status as significant covariates. Table 5 presents the final models, with missing estimates indicating that variable was not retained in the final model after backward selection (i.e., estimates with p≥0.05).

Table 5.

Final models presenting unstandardized beta coefficients for the associations of types of bullying experiences with body esteem and body dissatisfaction (n=328) separately for youth with obesity and youth without obesity^{^}.

Healthy & Overweight	Total Body Esteem (n=210)			Body Dissatisfaction (Boys only, n=99)			Body Dissatisfaction (Girls only, n=111)
	Beta	95% CIs	p-value	Beta	95% CIs	p-value	Beta	95% CIs	p-value
Bullying Sub scores
Cyber							−0.10	−0.17, −0.04	0.001*
Physical
Verbal-Relational	−0.32	−0.61, −0.03	0.033*
Covariates
Male
% of the 95th BMI Percentile	−0.27	−0.46, −0.09	0.005*	0.09	0.06, 0.11	<0.001*	0.06	0.04, 0.07	<0.001*
Age
Race
Parent Martial Status
In-school Status
Puberty Status
Household Quantity
	R squared = 0.06, F(2, 207) = 7.04			R squared = 0.36, F(1, 97) = 54.5			R squared = 0.32, F(2, 108) = 25.28
Obesity & Severe Obesity	Total Body Esteem (n=118)			Body Dissatisfaction (Boys only, n=50)			Body Dissatisfaction (Girls only, n=68)
	Beta	95% CIs	p-value	Beta	95% CIs	p-value	Beta	95% CIs	p-value
Bullying Sub scores
Cyber
Physical	1.23	0.11, 2.35	0.031*
Verbal-Relational	−0.99	−1.53, −0.46	<0.001*
Covariates
Male
% of the 95th BMI Percentile	−0.99	−1.53, −0.46	<0.001*	0.04	0.01, 0.06	0.003*	0.02	0.01, 0.03	0.001*
Age
Race
White	Reference						Reference
Black	9.63	3.12, 16.15	0.004*				−0.19	−0.77, 0.38	0.505
Other	−13.75	−35.83, 8.34	0.220				1.90	0.50, 3.30	0.009*
Parent Marital Status
Married	Reference
Divorced or Separated	6.06	−1.97, 14.08	0.138
Other	9.31	0.88, 17.73	0.031*
In-school Status
Puberty Status
Household Quantity
	R squared = 0.23, F(6, 111) = 5.47			R squared = 0.17, F(1, 48) = 10.17			R squared = 0.23, F(3, 64) = 6.33

^{^}Assessed using linear regression with backward selection retaining variables with p<0.10. Variables with missing estimates were not retained in the final model. The following covariates were entered into all models, with retained covariates displayed above: age, race, puberty status, in-school status, sex, percent of the 95^th BMI percentile, parent’s marital status, and household quantity.

^*p<0.05. CI = Confidence Interval.

Considering BD, percent of the 95^th BMI percentile was a consistent, positive predictor of BD across all sexes and weight status groups (p<0.05 for all models). Further, it was the only retained predictor in models with boys without obesity and with obesity. As for girls, cyberbullying experiences was negatively associated with BD among girls without obesity (n=111, B=−0.10, 95% CI= −0.17 to −0.04, p=0.001; R²=0.32), implying that greater cyberbullying victimization experiences resulted in less BD. No forms of bullying were associated with BD among adolescent girls with obesity (ps>0.05), only the covariates (race and percent of the 95^th BMI percentile) were significantly associated.

Discussion

The present study examined the relationships between bullying experiences, BE, and BD by obesity status. Bullying experiences were related to lower BE among adolescents with obesity as compared to adolescents without obesity. The relationships between total bullying experiences, total BE, and BD did not differ by sex; however, there were differences with bullying subscales predicting BD among girls but not boys. Differing types of bullying victimization contributed to BE and BD when evaluating adolescents with and without obesity separately, with some similarities (e.g., greater verbal-relational bullying experiences were associated with lower BE among both groups).

In line with the diathesis-stress model of bullying (Swearer & Hymel, 2015), the findings suggest that the BE of adolescents with obesity may be more impacted by bullying experiences as compared to adolescents without obesity. These findings align with a study of 43,000 youth (6^th to 12^th grade) which found that victimized youth with obesity, but not victimized youth with overweight, had greater levels of internalizing symptoms (e.g., anxiety, depression) compared to victimized youth with normal weight-status (Waasdorp et al., 2018), similarly suggesting that the impact of bullying experiences on cognitions may be heightened among youth with obesity. Furthermore, among adolescents with obesity in the present study, bullying sub-types were not related to BD, but verbal-relational and physical bullying experiences were associated with BE.

In this study, verbal-relational bullying was associated with BE, with the magnitude of the association greater among adolescents with obesity. This finding is similar to a previous study that found youth with obesity may be at greater risk of verbal bullying compared to normal weight youth (Wang et al., 2010); however, a dichotomous operationalization of victimization was used (compared to the current continuous frequency measure). Given the cross-sectional nature of the study, it may be that adolescents with low BE are at heighted risk for verbal-relational victimization, regardless of weight status.

The positive relationship between physical bullying victimization experiences and BE in adolescents with obesity was not found in other studies of physical bullying experiences (Garcia et al., 2019; Kukaswadia et al., 2011). One possible explanation is that adolescents who are physically bullied may themselves be perpetrators of physical bullying, proffering protections against low BE. For example, a power differential exists in bullying behavior between the perpetrator and the victim, and this could be from differences in size. Previous research has found greater overt (i.e., direct), as opposed to relational (i.e., indirect), perpetration among adolescents who reported a higher BMI (Schell-Busey et al., 2017). Youth who are physically bullied may recognize their size as a strength in a physical power differential situation, potentially elevating appreciation for one’s body, i.e., BE. This association was not found in relation to BD, which might be expected, as increased appreciation for one’s size would be more likely to be reflected in a global, attitudinal measure of body image than in perceptual judgments of current and ideal body size silhouettes, i.e., BD. Additional investigation into physical bullying and BE, including examination of perpetrators, victims, and bully-victims (Gower & Borowsky, 2013) may elucidate these relationships.

In line with previous findings that suggest sex and BMI differences in BD (Fernández-Bustos et al., 2019), the present study found that BD among girls, but not boys, was associated with specific bullying experiences. For adolescent girls without obesity, the present study found an inverse relationship between cyberbullying experiences and BD. Kenny and colleagues found that having at least one cyberbullying experience in the past couple of months was associated with greater BD among school-aged children (13 to 17 years of age) in Ireland (Kenny et al., 2018). Yet that study included children with unknown weight status and different measures of both cyberbullying and BD, which may explain the differences in findings compared to this study which separated by weight class. Those who report more cyberbullying may be closer to their reported “ideal” body type (i.e., BD scores around 0), and potentially more attuned to victimization on an aspect they are not interested in changing or feel they cannot change. The large sample of African American adolescents in this study may also reflect different cultural standards of body size (McCabe et al., 2013); however, race was not a significant predictor of BD in this subgroup of girls without obesity. It may be that adolescents without obesity are reporting negative BD scores (i.e., their current body image is thinner than their ideal body image), which may reflect a desire on the part of some adolescents—particularly males—to gain weight for athletic or aesthetic reasons (e.g., Wang, Yesalis & Fitzhugh, 1994). Together, the mechanisms amongst cyberbullying experiences and body image (i.e., lower BD) among only adolescents with overweight or normal weight warrant further examination.

Race was also significantly related to BD among adolescent girls with obesity, such that girls who identify as a race other than White or Black had higher BD as compared to adolescent girls with obesity who are White (B=1.90). These findings suggest that adolescent girls with obesity who are from racial groups other than Black or White may be particularly vulnerable to negative body image, which aligns with other findings demonstrating that negative body image is highest among Asian girls with obesity (Xanthopoulos et al., 2011). However, due to the small number of participants who identified as a race other than White or Black, these findings should be interpreted with caution. More research is needed to delineate this potential association, particularly given that students who are from racial/ethnic minority groups may be at heightened risk of victimization (Graham, 2006).

Finally, among adolescents with obesity, different forms of bullying victimization experiences were unrelated to BD but were related to BE. Victimization experiences may be most impactful on global evaluations of the self, i.e. BE, (as opposed to specific perceptions related to body size and shape ideals and how close or far their current body is to that ideal, i.e. BD) (Veron-Guidry & Williamson, 1996). BE also considers thoughts and feelings about one’s body in relation to others (e.g., “people my own age like my looks”), which may be more impacted by peer victimization, especially if the victimization is perceived to be weight-related than would someone’s perceptions of body ideal discrepancy. Previous research suggests that increased BD itself may be a predictor of victimization, independent of weight status (Sutter et al., 2015), so it may be expected that associations between BD and bullying victimization did not differ by weight status.

Strengths of the study include objective weight status measures, the use of two different measures of body image, and examination of frequency of and different types of bullying as they relate to self-evaluations. Other studies often use binary measures (presence/absence) of bullying experiences or do not examine several different types of bullying, which may have different impacts (Kenny et al., 2018). The current sample was composed of a larger portion of adolescents with obesity (36.0%) compared to similar studies,(Lee et al., 2018; Waasdorp et al., 2018) which enabled stratification by weight status.

Limitations

The cross-sectional study design of the present study limits the ability to draw causal conclusions. Previous research has prospectively linked bullying with increased weight among adolescents (Rosenthal et al., 2015; Schvey et al., 2019). Therefore, bullying itself may precede excess weight gain in adolescence. Longitudinal studies may provide additional explanation on the impact of bullying experience on BE and body image dissatisfaction by weight status. Additionally, longitudinal studies should aim to clarify the underlying mechanisms involved in these relationships, including potential mediators of the relationships between bullying experiences, body image and prospective weight gain, such as behavioral responses to body image disturbances and eating-related pathologies that contribute to increased adiposity. An additional limitation of the present study involves potential social desirability bias, although procedures were in place to ensure privacy during data collection including allowing the adolescent to complete the surveys in a private room on a laptop. Finally, the present study excluded adolescents with underweight, a group which warrants examination in future work.

Future Directions

The present measures may not be limited to bullying experiences from peers at school, as cyberbullying may be likely to occur outside of school. As such, it is unknown if participants reported on experiences outside of the school context (Bang et al., 2012; Puhl & Latner, 2007). Additionally, the present focus of the study was on bullying victimization, but accounting for the adolescent’s own involvement in bullying experiences by including bully-victims (i.e., those that engage in bullying behavior but are also victimized) is important to address, as adjustment problems can occur among perpetrators, victims, and bully-victims (Gower & Borowsky, 2018). Greater understanding of the interacting factors contributing to poor body image, with attention toward vulnerable groups such as adolescents with obesity who experience victimization and may be differentially affected by forms of bullying (Rettew & Pawlowski, 2016) can inform the tailoring of interventions aimed at improving coping skills, promoting positive body image, or reducing weight-bias (Diedrichs & Barlow, 2011). Last, the role of physical bullying victimization as a potential safeguard for their own low BE should be further explored.

Conclusion

Our study demonstrates the importance of weight status when considering the relationships between bullying experiences and body image, which differed by obesity status in adolescents. Additionally, this study contributes evidence that examining different types of bullying may provide greater nuance in uncovering the associations between bullying and body image. Attitudinal evaluations of the body, i.e., BE, may be more impacted by bullying experiences than perceptual types of body image, i.e., BD, which could have important implications for body image interventions. Our findings suggest that verbal-relational bullying victimization is associated with BE, especially amongst adolescents with obesity. Given the associations between bullying victimization, obesity, and body image, anti-bullying campaigns should address weight-related prejudices directed toward youth with obesity while fostering environments that promote acceptance of all body types.

Table 3.

Spearman rank correlations between bullying experiences, body esteem, and body dissatisfaction among adolescents (n=328).

Measures	1	2	3	4	5	6	7	8	9
1. Cyber Bullying
2. Physical Bullying	0.30***
3. Verbal-Relational Bullying	0.42***	0.65***
4. Total Bullying	0.50***	0.78***	0.96***
5. Appearance-related esteem	−0.20**	−0.16**	−0.28***	−0.25***
6. Attribution-related esteem	0.04	−0.10	−0.12*	−0.10	0.47***
7. Weight-related esteem	−0.11*	−0.04	−0.13*	−0.10	0.75***	0.47***
8. Total body esteem	−0.13*	−0.12*	−0.23***	−0.20***	0.92***	0.69***	0.88***
9. Body dissatisfaction	−0.02	0.01	0.08	0.05	−0.43***	−0.23***	−0.54***	−0.48***
10. Percent of 95th BMI Percentile	0.05	0.00	0.05	0.03	−0.34***	−0.20***	−0.54***	−0.42***	0.68***

BMI = Body Mass Index; Note:

^*indicates p<0.05

^**indicates p<0.01

^***indicates p<0.001

Abbreviations:

BD

body dissatisfaction

BE

body esteem

BIA

body image assessment

PECK

personal experiences checklist

Data availability statement:

Data collection for the study reported on in this paper is ongoing and therefore not archived. Data are available upon request to the corresponding author.