Children and Youth Services Review 104 (2019) 104385

Contents lists available at ScienceDirect

Children and Youth Services Review

journal homepage: www.elsevier.com/locate/childyouth

The indirect effect of bullying on adolescent self-rated health through T

mental health: A gender specific pattern☆
⁎
Anao Zhanga, , Chun Liub, Lindsay A. Bornheimera, Phyllis Solomonc, Kaipeng Wangd,
So'Phelia Morrowa
a
University of Michigan, School of Social Work, United States of America
b
The University of Texas at Austin, Steve Hicks School of Social Work, United States of America
c
University of Pennsylvania, School of Social Policy & Practice, United States of America
d
Texas State University, School of Social Work, United States of America

A R T I C LE I N FO A B S T R A C T

Keywords: Introduction: Bullying is a prevalent concern among adolescents and causes great mental/behavioral con-
Bullying sequences. In addition to the direct association between bullying victimization and adolescent physical health
Self-rated health (measured by self-rated health), this study investigates the indirect effect of mental health and the moderating
Mental health role of gender to such relationship.
Gender
Methods: Study participants comprised a national sample of adolescents (n = 3435) aged 15 to 16 years old in
Mediator
the United States.
Moderator
Results: Structural equation modeling analyses showed that, in addition to the main effect of bullying on ado-
lescents physical health, (based on joint significance test) mental health has an indirect effect on and gender
moderates the relationship between bullying victimization and adolescents' self-rated health.
Conclusions: Bullying has negative impact on adolescent health. Prevention and intervention programs need to
take into account the critical roles of gender and mental health in addressing the negative consequences of
bullying victimization.

1. Introduction (verbal) (National Center for Education Statistics, 2016; Wang, Iannotti,
& Nansel, 2009). More importantly, it is a significant social determinant
Adolescence is arguably one of the most significant yet challenging of health for this population (Halpern, Jutte, Colby, & Boyce, 2015;
developmental stages characterized by achieving numerous milestones, Viner et al., 2012) with recent research documenting the health con-
while dealing with rapidly unfolding emotional and social-relational sequences of being bullied as an adolescent (Jennings, Song, Kim,
incidents (Carr, 2015; Santrock, 2015). At this stage of development, it Fenimore, & Piquero, 2017; Zhang, Padilla, Kim Bendheim-Thoman,
is essential to have a healthy and successful completion of devel- 2018). Therefore, using data from the Fragile Family Wellbeing Studies,
opmentally required tasks, as healthy adolescent progression impacts the present investigation aims to (1) examine the relationship between
future biopsychosocial wellbeing (Sawyer et al., 2012). Adverse events bullying victimization and adolescent general health status; (2) explore
within the social environments of adolescents, such as bullying, can the potential indirect effect of mental health in the relationship be-
derail a positive developmental trajectory. Bullying is defined as “un- tween bullying and adolescent health; and (3) explore a possible
wanted, aggressive behavior among school-aged children that involves moderating role of sex in the relationships between adolescent health,
a real or perceived power imbalance. The behavior is repeated, or has bullying, and mental health.
the potential to be repeated, over time” (US Department of Health and
Human Services, 2014). Bullying, whether physical, verbal, or emo- 2. The importance of and challenges to adolescent health
tional, can negatively impact the health of adolescents and its effects
can continue into adulthood. Among adolescents, bullying occurs at an Positive adolescent general health is associated with lower risk of
overall 30% prevalence rate, ranging from 20.8% (physical) to 54% substance abuse (Haberstick et al., 2014), improved cognitive function

☆
The work was primarily conducted at the University of Michigan, School of Social Work.
⁎
Corresponding author at: University of Michigan, School of Social Work, RM 3704, 1080 S. University Ave, Ann Arbor, MI 48109, United States of America.
E-mail address: [email protected] (A. Zhang).

https://doi.org/10.1016/j.childyouth.2019.104385
Received 14 March 2019; Received in revised form 17 June 2019
Available online 20 June 2019
0190-7409/ © 2019 Elsevier Ltd. All rights reserved.
A. Zhang, et al. Children and Youth Services Review 104 (2019) 104385

(Verburgh, Königs, Scherder, & Oosterlaan, 2014), reduced mental other objective health measures like BMI or physician's diagnosis. First,
health challenges (McLeod, Horwood, & Fergusson, 2016), and heal- self-rated health is a more inclusive measure of health status. When
thier BMI scores (Schofield, Conger, Gonzales, & Merrick, 2016) as asked about self-rated health, individuals often include both objective
adolescents transition into young adulthood (Santrock, 2015). At this health characteristics as well as their subjective evaluation of health,
high-stakes stage, adolescents encounter enormous challenges across all thus presenting a more comprehensive picture than most additive
bio-psycho-social levels (Catalano et al., 2012). Processing the physical measures of disease (Idler & Benyamini, 1997). Second, self-rated
and emotional changes of puberty, adolescents are faced with a fast- health is a dynamic measure of one's health as it incorporates one's on-
maturing brain as well as numerous peaks of neurological, cognitive going surveillance of his/her health which contains a wealth of in-
and emotional developments (Spear, 2013; Steinberg, 2005). Simulta- formation, including mental health status, that are often inaccessible to
neously, adolescents are confronted with the urgent need to form self- a health care provider or not provided by another health measure (e.g.,
identity, which subsequently results in numerous psychosocial chal- Benyamini, 2011). Finally, self-rated health reflects psycho-social re-
lenges to adolescents' development (Newman & Newman, 2017). Over sources and associations that are beyond biological indicators of one's
the past two decades, literature has reached a consensus on including health, such as education, social support, access to health services, and
social and contextual patterns as factors affecting adolescent health, other socio-economic factors. All these factors are stronger predictors of
often termed social determinants of adolescent health (Currie et al., individual's future health status and disease trajectory. Therefore, this
2009; Viner et al., 2012). Peer relationships, for example, are salient study used self-rated health as the dependent variable to measure
proximal social determinants of adolescent health with research in- adolescents' general health status.
dicating associations between positive peer relationships and higher
self-esteem (Shroff & Thompson, 2006), fewer behavioral problems 4. Theoretical framework
(Rambaran et al., 2017), and greater life satisfaction (Brière, Pascal,
Dupéré, & Janosz, 2013). We primarily used Diez Roux's (2012) pathway model of health
disparity in guiding the conceptualization of our study. The pathway
3. Bullying as an influential social determinant for Adolescents' model articulates that environmental context and life course factors are
health integral to understanding health disparity. For example, family (Moore
et al., 2018) and social environmental context (Fong, Cruwys, Haslam,
Bullying is arguably one of the most significant social determinants & Haslam, 2019) have become increasingly significant contributors to
of adolescents' general health status. Preoccupied by developmental individuals' health. For the adolescent population, relationship with
tasks, adolescents focus a majority of their effort on exploring and so- peers is arguably a key social environmental factor for adolescents'
lidifying their self-identity, largely through peer interactions. The health (e.g., Williams & Anthony, 2015). Therefore, it is expected that
quality of these interactions are essential to a healthy developmental positive or negative relationship with peers is likely to affect adoles-
process, while adversity and distress arising from these relationships cents' health.
can substantially impact overall wellbeing (Oberst, Wegmann, Stodt, Building on a wealth of literature of social determinants of health,
Brand, & Chamarro, 2017; Pouwels, Scholte, van Noorden, Tirza, & the pathway model further emphasizes the importance of studying
Cillessen, 2016). Bullying has become a common concern among school potential mechanism through which environmental factors affect ado-
age children and adolescents (Craig et al., 2009). Unlike quarrels or lescents' health status (Diez Roux, 2012). One often studies mechanisms
physical fights, bullying is a unique form of social and relational hos- that involve psychological factors, such as distress, depression or low
tility that repeatedly exposes individuals to an imbalanced power dy- self-esteem (Balázs et al., 2013; Duchesne et al., 2017). Specifically, a
namic (Jimerson & Swearer, 2009). Exposure to such hostility induces longitudinal cohort study identified self-esteem as being a significant
vigilance among adolescents and creates harmful health effects through predictor for adolescents' self-rated health (Jerdén, Burell, Stenlund,
intermediary processes, such as negative emotions or psychological Weinehall, & Bergström, 2011). Similarly, another longitudinal study
distress, which over time may result in chronic stress (Roland, 2002; found adolescents' depression at age 15 significantly predicted adoles-
Sigurdson, Undheim, Wallander, Lydersen, & Sund, 2015; Thomas, cents' physical health status at age 20 (Keenan-Miller, Hammen, &
Connor, & Scott, 2018). Through neuroendocrine pathways, chronic Brennan, 2007). Guided by the pathway model of health disparity, the
stress weakens adolescents' immune system and can result in physical current study investigated whether social factors (bullying victimiza-
health problems (Suárez, Feramisco, Koo, & Steinhoff, 2012). tion) affect adolescent self-rated health, as well as the possible me-
A compelling body of empirical literature has linked exposure to chanisms through which these factors impact self-rated health.
bullying and adolescent general health status. Gruber and Fineran
(2008), for example, found adolescents who experience bullying have 5. Possible indirect effect of mental health in relationship
poorer symptoms of physical health, such as sleeping disturbances, between bullying and self-rated health
upset stomachs, migraines or dizziness than those who did not experi-
ence bullying. Other research has determined that the experience of With increasing empirical evidence of the association between
bullying relates to higher risk of bedwetting, fatigue, and poor appetite bullying and adolescent general health status, research has suggested
(Fekkes, Pijpers, Fredriks, Vogels, & Verloove-Vanhorick, 2006; mental health having an indirect effect in this relationship. Fekkes,
Kumpulainen & Räsänen, 2000; Løhre, Lydersen, Paulsen, Mæhle, & Pijpers, and Verloove-Vanhorick (2004) found bullying to be associated
Vatten, 2011). A meta-analysis of 11 studies reported that adolescents with both depression and psychosomatic complaints. The authors
who experienced bullying were significantly more likely to develop commented that bullying may not directly impact health problems, but
somatic symptoms (Gini & Pozzoli, 2009). Furthermore, lower exposure the relationship between psychological distress (e.g. depression) and
to bullying has been observed to be significantly associated with higher bullying may produce a harmful impact on adolescents' general health
levels of self-assessed health among children from 9 to 14 years old status. Another study (Fekkes et al., 2006) examined the temporal or-
(Danese & Lewis, 2017; Forrest, Bevans, Riley, Crespo, & Louis, 2013). dering of bullying victimization and adolescent physical wellbeing and
To assess adolescents' general health, this study used a well-estab- found that a sequela of a bullying incident frequently resulted in mental
lished measure: self-rated health. Self-rated health is a single-item health symptoms of depression and anxiety, which were subsequently
measure which has received strong support both from conceptual and followed by various health symptoms. This ordering sequence supports
empirical literature (e.g., Jylhä, 2009; Zhang, De Luca, Oh, Liu, & Song, the potential indirect effect of mental health symptoms in the re-
2019). Conceptual studies have identified reasons why self-rated health lationship between bullying and general health status including phy-
is a stronger proxy for adolescents' general health status than most sical health of adolescents. While investigators have suggested a

2
A. Zhang, et al. Children and Youth Services Review 104 (2019) 104385

possible indirect effect of mental health in this relationship, to our 8. Methods

knowledge, none to date have conducted such an analysis.
8.1. Data and sample

6. Moderating effects of sex This study used the most recent wave (year 15 follow-up) of data
from the Fragile Families and Child Wellbeing Studies (FFCWS)
Both bullying and adolescent health literature have emphasized the (Reichman, Teitler, Garfinkel, & McLanahan, 2001), which was com-
importance of considering sex differences in understanding the health prised at baseline of 4898 infants born between 1998 and 2000. Par-
consequences of a bullying experience. For example, male adolescents ticipants (focal children and their parents) were drawn from a stratified
are more likely to be victims of bullying, as limited findings suggest sample of 20 major cities in the United States with populations of
they have a greater likelihood of being excluded or rejected by peers 200,000 or more. The sample includes adolescents in both married and
than their female counterparts (Veenstra, Lindenberg, Munniksma, & unmarried families, with an over-representation from unmarried fa-
Dijkstra, 2010). In addition, sex differences are present in the nature of milies, i.e., “fragile families”. In prior waves, FFCWS interviewed the
bullying experienced. Male adolescents more commonly encounter focal children's primary caregivers (mostly mothers) and their other
physical and verbal aggression from their peers (e.g., hitting, kicking or biological parent (who may or may not be the primary caregiver). In its
cursing), while female adolescents more often experience relational most recent wave (collected from 2014 to 2017), the study includes
exclusion (e.g. being socially excluded) (Orpinas, McNicholas, & data from child participants who are currently 15–16 years old and
Nahapetyan, 2015; Stubbs-Richardson et al. Bendheim-Thoman, 2018). their primary caregiver (PCG) from separate interviews conducted via
Moreover, females and males differ in how they cope with bullying phone. The year 15 follow-up survey contains 3580 PCG (77% of the
incidents and its consequences. Research has determined greater help- baseline sample) and 3444 adolescents (74% of the baseline sample)
seeking behavior on the part of female adolescents, while males tend to with each completing a 1-h telephone interview (Bendheim-Thoman
use more passive coping strategies in response to bullying (Dukes, Stein, Bendheim-Thoman, 2018). For purposes of the present analyses, the
& Zane, 2010; Machmutow, Perren, Sticca, & Alsaker, 2012). Conse- sample consisted of 3435 adolescents (predominantly 15 or 16 years
quently, to understand the relationship between bullying victimization, old) who reported complete data on their self-rated health.
adolescent health, and mental health, it is important to consider a
possible moderating role of sex. 8.2. Measures
Furthermore, moderating effects of sex may extend beyond the
bullying–adolescent health relationship by being associated with an 8.2.1. Dependent variable
indirect effect of mental health on this relationship. Prior research has Self-rated health. Adolescents' physical health was measured by a
consistently reported that sex plays a critical role in differentiating global measure of health (Bauldry, Shanahan, Boardman, Miech, &
mental health consequences of bullying. Bakker, Ormel, Verhulst, and Macmillan, 2012; Zhang, Padilla, & Kim, 2017). Adolescents were
Oldehinkel (2010) tested sex-specific mental health effects of peer asked “overall, how would you rate your health?” Participants re-
stressors during adolescence and found relationship losses (due to sponded using a 5-point Likert scale ranging from “1 = Poor” to “5 =
bullying) were more strongly associated with mental health problems in Excellent”, with a higher score indicating a better self-rated health.
girls than boys. Another study (Turner, Exum, Brame, & Holt, 2013)
examined the differential effects of bullying victimization by sex and 8.2.2. Focal variable
concluded sex was a critical component in understanding adolescents' Bullying Victimization. Bullying victimization was measured by the
mental health consequences of bullying. frequency of how often in the last month the adolescents experienced
bullying behaviors including (a) “picked on me or said mean things to
me”, (b) “hit me”, (c) “took my things”, or (d) “purposely left me out of
7. Research hypotheses activities”. An adolescent reported from “never = 0” to “about every
day = 4” for each of the four forms of bullying, and the responses were
While previous research has provided a strong empirical foundation summed to create an index of bullying with scores ranging from 0 to 16,
for evaluating the relationship between bullying, adolescents' mental with a higher score indicating higher degree of bullying victimization.
health, self-rated health and the possible moderating role of sex to such Adolescent Depression and Anxiety. Adolescent depression was
relationship, few studies to our knowledge have pursued these factors measured by 5 items from the Center for Epidemiologic Studies
simultaneously. Furthermore, most previous studies used school-based Depression Scale (CES-D) (Radloff, 1977) and anxiety was measured by
samples with limited generalizability. Therefore, using a national 6 items from the Brief Symptom Inventory 18 (BSI 18) anxiety subscale
sample, this study investigated the relationship between bullying, (Derogatis and Savitz, 1999). For adolescent depression, the 5 items
adolescents' mental health and self-rated health as well as considered were selected based on previous psychometric studies suggesting these
the indirect effect of mental health and sex as a moderator. Study hy- items to be an improvement over the full 20-item CES-D in cross-cul-
potheses include: (1) greater experience of bullying victimization will tural comparability (Perreira, Deeb-Sossa, Harris, & Boleen, 2005).
be independently associated with poorer adolescent mental health and Adolescents responded to a 4-point Likert scale from “Not at all = 1” to
self-rated health among the full sample (1a) and independently among “Extremely = 4” with a higher score representing a higher degree of
males and females (1b); (2) adolescent mental health will have an in- depression or anxiety symptoms (ranging from 5 to 20 and 6 to 24 for
direct effect on the relationship between bullying and physical health depression and anxiety, respectively). Both subscales had satisfactory
among the full sample (2a) and independently among males and fe- internal consistency with Cronbach's alpha of 0.97 and 0.98 for de-
males (2b). Specifically, greater experience of bullying will be asso- pression and anxiety subscales, respectively.
ciated with poorer mental health and, in turn, associated with poorer Sex. Adolescent sex was that assigned at birth with male = 1, fe-
self-rated physical health; and (3) sex will moderate the relationships male = 2.
between bullying, mental health and self-rated health. We therefore
hypothesize the indirect effect of mental health on the relationship 8.2.3. Covariates
between bullying and self-rated health will be stronger among females Informed by Diez Roux's (2012) pathway model of health disparity,
than males. this study also controlled for other known correlates of bullying victi-
mization and self-rated health, including: demographic variables, fa-
mily socioeconomic status, adolescents' socially supportive

3
A. Zhang, et al. Children and Youth Services Review 104 (2019) 104385

Table 1
Participants' characteristics (N = 3435).
Totala Maleb Femalec Diffd

Bullying victimization 0.70 / 1.67 0.71 / 1.72 0.68 / 1.63 p = .636

Depression 8.00 / 2.99 7.71 / 2.81 8.31 / 3.15 p < .001
Anxiety 10.88 / 3.91 10.61 / 3.67 11.16 / 4.13 p < .001
Sex (female) 1664 (48.4%)
Race (White) 590 (18.1%) 298 (17.9%) 292 (18.3%) p = .061
Black 1599 (49.0%) 805 (48.3%) 794 (49.8%)
Hispanic 813 (24.9%) 423 (25.4%) 390 (24.5%)
Other only 86 (2.6%) 57 (3.4%) 29 (1.8%)
Multi-racial 174 (5.3%) 85 (5.1%) 89 (5.6%)
PCG Relationship (married / living together) 938 (31.2%) 504 (32.8%) 434 (29.6%) p = .211
Involved but not living together 530 (17.6%) 265 (17.3%) 265 (18.1%)
Separated, divorced or widow 584 (19.4%) 283 (18.4%) 301 (20.5%)
No relationship 952 (31.7%) 484 (18.4%) 468 (20.5%) p = .362
PCG Education (< high school) 600 (17.7%) 294 (16.8%) 306 (18.6%)
High school or equivalent 670 (19.7%) 350 (20.0%) 320 (19.4%)
Some college 1485 (43.7%) 759 (43.4%) 726 (44.0%)
College degree or higher 644 (18.9%) 347 (19.8%) 297 (18.0%)
Family income per capita (< 19,999) 725 (21.2%) 352 (20.0%) 373 (22.6%) p = .226
20,000–39,999 864 (25.3) 442 (25.1%) 422 (25.5%)
40,000–74,999 901 (26.4) 475 (26.9%) 426 (25.8%)
75,000 or higher 926 (27.1) 494 (28.0%) 432 (26.1%)
School climate 31.49 / 5.16 31.74 / 5.01 31.23 / 5.30 p < 0.01
Diagnosed physical problems 0.27 / 0.55 0.24 / 0.53 0.28 / 0.56 p = .06

a
Total sample descriptive statistics: mean and standard deviation (mean / SD) for continuous variable, frequency and percentage (n (%)) for categorical variable.
b
Male subsample descriptive statistics.
c
Female subsample descriptive statistics.
d
P values indicating the difference between male and female subsamples descriptive statistics, t-test for continuous variable and chi-square test for categorical
variables.

environment (school climate) and adolescents' diagnosed physical variable normality and missingness. Skewness was in the range of ± 2
health conditions. Adolescents' race/ethnicity (non-Hispanic white = 1, and kurtosis ± 7 for all variables. Data missingness ranged from 0% to
non-Hispanic black = 2, Hispanic = 3, other only, non-Hispanic = 4, 8.6% and a general missing data pattern was not observed. Results of
and multi-racial, non-Hispanic = 5). the Little's MCAR test (Little, 1988) suggested the mechanism of miss-
Primary Caregiver (PCG) education was coded as less than high ingness as completely at random, χ2(86) = 42.35, p > .05. Pre-
school = 1, high school or equivalent = 2, some college = 3, college/ liminary analyses in SPSS involved univariate and bivariate examina-
graduate degree or above = 4. PCG's relationship status with the other tion of all variables using crosstabs with chi-square tests for categorical
biological parent was measured as married or living together = 1, ro- and independent samples t-tests for continuous variables. Hypothesis
mantically involved or friends but not living together = 2, separated, testing utilized Structural Equation Modeling (SEM) in Mplus 8 with a
divorced, or widowed = 3, and no relationship at all or unknown bio- robust (Huber-White) maximum likelihood algorithm for non-normality
logical parent = 4. Family socioeconomic status was measured by re- and variance heterogeneity. Missing data, though minimal and com-
ported annual household income. This variable was grouped into pletely at random (see data preparation above), were analyzed using
1 ≤ 19,999, 2 = 20,000–39,999, 3 = 40,000–74,999, and 4 = 75,000 Full Information Maximum Likelihood (FIML) methods; and coefficients
or higher. alphas/composite reliabilities were evaluated for multi-item measures.
School climate was measured by a 10-item scale (see Measures of In Model 1, adolescent self-rated health served as the endogenous
Effective Teaching [MET] Project; Kane & Cantrell, 2010) reflecting variable and the latent mental health variable was specified to re-
adolescents' perceptions of their teachers' and peers' emotional support. present the common variance of anxiety and depression. The experience
Adolescents responded to questions such as “The teachers in this school of bullying victimization represented the exogenous variable and cov-
treat the students with respect” or “Kids in this school behave the way ariates included sex, race, Primary caregiver (PCG) relationship status,
the teachers want them to” on a 4-point Likert scale, i.e., strongly PCG education, family income, adolescent reported school climate and
agree = 1 to strongly disagree = 4. The responses were reverse coded adolescents diagnosed physical health problems. The latent mental
and then summed to form an index of overall school climate, ranging health construct was examined for its indirect effect in the relationship
from 4 to 40 so that higher scores represented a more supportive school between bullying and self-rated health (hypothesis 2). In Model 2, sex
climate. The measure had satisfactory internal consistency with Cron- was examined as a moderator in the relationships between bullying and
bach's Alpha = 0.86. self-rated health, bullying and mental health and self-rated health
Adolescents' diagnosed with physical health problems was mea- (hypothesis 3), using multiple group analysis with the ‘Grouping’ com-
sured by parental report to the question “has a doctor or health pro- mand in Mplus. The fit of the SEM model was evaluated using both
fessional ever told you that (your child) has any of the following con- global and focused fit indices (Muthén & Muthén, 2013). Global fit
ditions?” Parents selected 1 = yes or 0 = no for a list of 7 physical using chi-square (χ2 > 0.05 or χ2/df ratio < 3.0), root mean square
health diagnoses (e.g., asthma, anemia, congenital heart disease). The error of approximation (RMSEA < 0.08 (90% CI)), standardized root
summed scores ranged between 0 and 7. mean square residual (SRMR < 0.05), comparative fit index (CFI >
0.95), Tucker Lewis Index (TLI) and focused fit using standardized
residuals (< |2|) and modification indices (< |4|). All statistical models
9. Data modeling and analysis were unweighted. However, the coefficients and standard errors should
not be affected because we controlled for primary caregivers' char-
Data were analyzed in SPSS24 (IBM Corp, 2016) and Mplus 8 acteristics that were used in creating weights for the Fragile Families
(Muthén & Muthén, 2013). Data preparation included examinations of

4
A. Zhang, et al. Children and Youth Services Review 104 (2019) 104385

data in all of the analytic models, which is the common approach

employed in previous studies using this dataset (e.g., Kim, Padilla,
Zhang, & Oh, 2018; Schmeer, 2012; Zhang et al., 2019).

10. Results

Characteristics of participants are presented in Table 1. About half

of the study participants were female (n = 1664, 48.4%) and almost
half of them were black (n = 1599, 49%). Over 30% of the primary
caregivers were married or living together with the other biological
parent (n = 938, 31.2%) and close to half of the primary caregivers had
some college education (n = 1485, 43.7%). Adolescent participants
reported an average score of 8 (out of a possible of 20) and 11 (out of
possible of 24) for depression and anxiety, respectively. Participants
reported an average of 31.49 (out of 40) for school climate, indicating
an overall supportive school climate. An overall score lower than one
(mean = 0.27) was reported for diagnosed physical health problems by
Fig. 1. Model 1 relationships among both males and females.
PCG, indicating few physical health problems. Significance between sex Notes: Standardized coefficients are represented first, followed by “/” and un-
differences was observed, with female adolescents reporting sig- standardized coefficients; *p < .001.
nificantly greater levels of depression (mean = 7.71 (male) and 8.31
(female), t(3382) = −5.87, p < .001), anxiety (mean = 10.61 (male)
unit increase in mental health, there was an average associated 0.25
and 11.16 (female), t(3362) = −4.13, p < .001), and lower levels of
unit decrease in self-rated health (Standard Error (SE) = 0.02, Critical
supportive school climate (mean = 31.74 (male) and 31.23 (female), t
Ratio (CR) = −11.0, p < 0.001 (standardized) / b = −0.08,
(3361) = −2.91, p < .01).
SE = 0.01, CR = −9.53, p < .001 (unstandardized)), holding bullying
and all covariates constant.
10.1. Model 1 Furthermore, the relationship between bullying and self-rated
health was indirectly related through mental health as indicated by the
All global fit indices indicated good model fit (χ2 = 4.46, p = .73, joint significance test (MacKinnon, Lockwood, Hoffman, West, &
χ /df = 0.64,
2
CFI = 0.998, RMSEA = 0.004 (0.001–0.013), Sheets, 2002). For every one-unit increase in bullying, there was an
TLI = 1.00, and SRMR = 0.003). Examination of focused fit indices average associated 0.31 unit increase in severity of mental health
(standardized residuals and modification indices) revealed no theore- (SE = 0.03, CR = 1.11, p < 0.001) / b = 0.54, SE = 0.06, CR = 9.29,
tically meaningful points of stress on the model. Parameter estimates p < .001), holding self-rated health and covariates constant. Taken
are reported in Table 2 and Fig. 1 presents the standardized and un- together, the experience of bullying victimization related to poorer
standardized parameter estimates for the structural and measurement adolescent self-rated health indirectly through mental health (Sobel
model among the full sample (both males and females). test = 7.96, p < .001), in which greater bullying related to poorer
Bullying victimization, mental health, and all covariates accounted mental health, and, subsequently poorer mental health related to
for 12% of the variance in self-rated health for the full sample (both poorer self-rated health. The total effect of bullying and self-rated
males and females). Findings indicated that adolescent mental health health was also investigated given there existed an indirect effect of
significantly related to self-rated health (hypothesis 1a). For every one-

Table 2
Parameter estimates for paths in models 1 and 2.
Pathway Model 1 Model 2

Male Female

b SE CR p b SE CR p b SE CR p

Measurement modela
⁎ ⁎ ⁎
Mental health by anxiety 0.70 0.02 35.81 0.72 0.02 34.16 0.74 0.02 36.45
⁎ ⁎ ⁎
Mental health by depression 0.96 0.03 34.05 0.88 0.02 36.47 0.91 0.02 44.51
Structural modela
⁎ ⁎ ⁎
Self-rated health on mental health −0.25 0.02 −11.0 −0.22 0.03 −7.42 −0.26 0.03 −8.16
⁎
Self-rated health on bullying −0.03 0.02 −1.30 −0.11 0.03 −3.62 0.02 0.03 0.56
⁎ ⁎ ⁎
Mental health on bullying 0.31 0.03 11.11 0.24 0.03 7.57 0.27 0.03 8.79
Measurement modelb
⁎ ⁎
Mental health by anxiety 0.96 0.05 17.72 ⁎ 1.06 0.05 20.66 1.06 0.05 20.66
Mental health by depression 1.00 0.00 999 1.00 0.00 999 1.00 0.00 999
Structural modelb
⁎ ⁎ ⁎
Self-rated health on mental health −0.08 0.01 −9.53 −0.08 0.02 −6.94 −0.09 0.02 −7.59
⁎
Self-rated health on bullying −0.02 0.01 −1.30 −0.06 0.02 −3.72 0.01 0.02 0.56
⁎ ⁎ ⁎
Mental health on bullying 0.54 0.06 9.29 0.35 0.05 7.11 0.47 0.05 9.55

Global fit indices: χ2 = 4.46, p = .73, χ2/df = 0.64, CFI = 0.998, RMSEA = 0.004 (0.001–0.013), TLI = 1.00, and SRMR = 0.003 (Model 1), χ2 = 33.62, p = .07,
χ2/df = 2.10, CFI = 0.992, RMSEA = 0.025 (0.013–0.036), TLI = 0.977, and SRMR = 0.008 (Model 2).
SE = standard error; CR = critical ratio.
⁎
p < .001.
a
Standardized coefficients.
b
Unstandardized coefficients.

5
A. Zhang, et al. Children and Youth Services Review 104 (2019) 104385

Fig. 2. Model 2 relationships among males. Fig. 3. Model 2 relationships among females.
Notes: Standardized coefficients are represented first, followed by “/” and un- Notes: Standardized coefficients are represented first, followed by “/” and un-
standardized coefficients; *p < .001. standardized coefficients; *p < .001.

mental health to this relationship. For every-one unit increase in bul- 10.2.2. Females
lying, on average, there was an associated 0.11 unit decrease in self- Fig. 3 presents the standardized and unstandardized parameter es-
rated health when holding mental health and covariates constant. timates for the structural and measurement model with standard error
in parentheses for females. Bullying victimization, mental health, and
all covariates accounted for 10% of the variance in self-rated health for
10.2. Model 2 females. Findings indicated that adolescent mental health significantly
related to self-rated health, yet bullying victimization did not directly
All global fit indices indicated good model fit (χ2 = 33.62, p = .07, relate to self-rated health (hypothesis 1b). Bullying, however, did relate
χ /df = 2.10,
2
CFI = 0.992, RMSEA = 0.025 (0.013–0.036), significantly to self-rated health through mental health (hypothesis 2b)
TLI = 0.977, and SRMR = 0.008). Examination of focused fit indices as indicated by the joint significance test (MacKinnon et al., 2002). For
(standardized residuals and modification indices) revealed no theore- every one-unit increase in bullying, there was an average associated
tically meaningful points of stress on the model. 0.27 unit increase in mental health (SE = 0.03, CR = 8.79, p < .001 /
b = 0.47, SE = 0.05, CR = 9.55, p < .001), holding self-rated health
and covariates constant. For every one-unit increase in mental health,
10.2.1. Males there was an average associated 0.26 unit decrease in self-rated health
Fig. 2 presents the standardized and unstandardized parameter es- (SE = 0.03, CR = −8.16, p < 0.001 / b = −0.09, SE = 0.02,
timates for the structural and measurement model with standard error CR = −7.59, p < .001), holding bullying and covariates constant.
in parentheses for males. Bullying victimization, mental health, and all Taken together, the experience of bullying victimization for females
covariates accounted for 11% of the variance in self-rated health for related to poorer adolescent self-rated health indirectly through mental
males. Findings indicated that adolescent mental health and bullying health (Sobel test = −6.24, p < .001), in which greater bullying re-
significantly related to self-rated health, independently (hypothesis 1b). lated to poorer mental health, and, subsequently poorer mental health
For every one-unit increase in bullying, there was an average associated related to poorer self-rated health. The total effect of bullying and self-
0.11 unit decrease in self-rated health (SE = 0.03, CR = −3.67, rated health was also investigated given there existed an indirect effect
p < 0.001) / b = −0.06, SE = 0.02, CR = −3.72, p < .001), holding of mental health to this relationship. For every-one unit increase in
mental health and covariates constant. bullying, there was an average associated 0.05 unit decrease in self-
Furthermore, the relationship between bullying and self-rated rated health when holding mental health and covariates constant
health were both directly and indirectly related through mental health among females.
as indicated by the joint significance test (MacKinnon et al., 2002). For
every one-unit increase in bullying, there was an average associated
0.24 unit increase in mental health (SE = 0.03, CR = 7.57, p < 0.001) 10.2.3. Moderating effects of sex
/ b = 0.35, SE = 0.05, CR = 7.11, p < .001), holding self-rated health Pertaining to hypothesis 3, path coefficients were compared for both
and covariates constant. For every one-unit increase in mental health, groups (male and female) in a multiple group analysis by using a critical
there was an average associated 0.22 unit decrease in self-rated health ratio over 1.96 (over 95% confidence). Thus, it was concluded that
(SE = 0.03, CR = −7.42, p < 0.001) / b = −0.08, SE = 0.02, paths were moderated by sex if the critical ratio of the coefficient was
CR = −6.94, p < .001), holding bullying and covariates constant. over 1.96 (Fairchild & MacKinnon, 2009; Noor-Azniza, Malek, Yahya, &
Taken together, the experience of bullying victimization for males Farid, 2011). As can be seen in comparing pathways of Figs. 1 and 2, a
related to poorer adolescent self-rated health both directly and in- significant moderating effect for sex was present in the relationship
directly through mental health (Sobel test = −0.05, p < .001), in between bullying and self-rated health (b = − 0.07, SE = 0.02,
which a greater degree of bullying related to poorer mental health, and CR = −2.95, p < .001). Essentially, the relationship between bullying
subsequently poorer mental health related to poorer self-rated health. victimization and adolescent self-rated health varied as a function of
The total effect of bullying and self-rated health was also investigated sex, with bullying relating to self-rated health only for males
given there existed an indirect effect of mental health to this relation- (b = −0.11, SE = 0.03, CR = 3.67, p < .001 / b = −0.06, SE = 0.02,
ship. For every-one unit increase in bullying, there was an average CR = −3.72, p < .001) and not for females (b = 0.02, SE = 0.03,
associated 0.16 unit decrease in self-rated health when holding mental CR = 0.56, p > .05 / b = 0.01, SE = 0.02, CR = 0.56, p > .05). There
health and covariates constant among males. was no moderating effect found for sex in the relationship between

6
A. Zhang, et al. Children and Youth Services Review 104 (2019) 104385

mental health and self-rated health, yet, a trend was apparent in the understanding the harmful effects of bullying victimization for adoles-
relationship between bullying and mental health (b = −0.13, cent health by sex. Furthermore, this finding also indicates that there
SE = 0.07, CR = 1.84, p = .06). Thus, the relationship between bul- are greater mental health impacts as a consequence of bullying among
lying victimization and mental health tended toward significantly female than their male counterparts. This reinforces an already com-
(p = .06) varying as a function of one's assigned sex. The relationships pelling body of literature on the differential mental health effect by sex.
between bullying and mental health were significant for both males and For example, Turner, Finkelhor, and Ormrod (2006) reported that fe-
females, with females experiencing a greater increase in mental health male adolescents more often internalize their victimization experience
symptoms (b = 0.27, SE = 0.03, CR = 8.79, p < .001 / b = 0.47, (e.g., depression, anxiety), whereas male adolescents tend to engage in
SE = 0.05, CR = 9.55, p < .001) as compared to their male counter- externalizing behaviors (e.g., aggression) as a consequence. Con-
parts (b = 0.24, SE = 0.03, CR = 7.57, p < 0.001 / b = 0.35, sidering that the latent mental health variable was constructed from
SE = 0.05, CR = 7.11, p < .001). internalizing, i.e., depression and anxiety scales, rather than measuring
externalizing behaviors, the stronger association among female ado-
11. Discussion lescents was not unexpected. Future research should measure both in-
ternalizing and externalizing behaviors.
Building on a fast-growing body of literature of the relationship The study has some significant practice implications for health and
between bullying victimization and adolescent health, this study con- mental health professionals working with adolescents especially in
firms the harmful consequences of bullying experience on adolescent schools. Given the critical role of mental health in the healthy devel-
self-rated health – a global measure of adolescent general health status opment of adolescents, and the requisite influence of bullying as a so-
(Boardman, 2006; Zhang et al., 2017; Zhang et al., 2019). More im- cial determinant for adolescents' health, preventive programs targeting
portantly, results of this study further advance the literature by em- a reduction in bullying in school or community settings are imperative.
pirically demonstrating the indirect effect of mental health and the Additionally, when treating adolescents with health complaints, school-
moderating role of sex in the relationship between bullying victimiza- based health and mental health professionals, specifically school
tion and adolescent self-rated health. Both theoretical and empirical nurses, should assess for mental health symptoms and risk factors of
literature has supported the interconnection between mental and phy- bullying victimization. Further, after an episode of bullying, even if
sical health over the past decades (Danese & Lewis, 2017; Kemp & only the presence of a physical encounter with no verbal exchange, it is
Quintana, 2013). Given the consensus in the extant literature that critical to assess for mental health symptoms. Finally, given the sex-
bullying creates significant mental health challenges among adolescents specific pattern between bullying victimization and health, tailored
(e.g., psychological distress, major depression, and severe anxiety), its services and programs need to be made available for adolescents, taking
social impact on adolescent physical health is expected (Fekkes et al., into account potential sex differences. For example, female adolescents
2006; Kumpulainen & Räsänen, 2000; Løhre et al., 2011). may benefit more from an integration of mental health treatment into
Moving beyond the direct association between bullying and self- physical health services, while their male counterparts may benefit
rated health, the present study found mental health has an indirect more from physical health services being integrated to mental health
effect to the bullying and self-rated health relationship with bullying treatment.
victimization being associated with lower adolescent self-rated health Several limitations should be noted for the present study. First, the
through poorer mental health. Among males, the experience of bullying results should not be construed as a causal relationship but as asso-
victimization was related to poorer self-rated health both directly and ciational. However, by identifying mental health's indirect effect on the
indirectly through mental health. Among females, the experience of relationship between bullying victimization and adolescents' self-rated
bullying victimization was related with poorer self-rated health in- health, it strengthens the evidence for a possible causal association
directly through mental health. Thus, there was no direct relationship between bullying and self-rated health. Second, this study employed
between bullying and self-rated health among female adolescents, self-rated health as a global measure of adolescents' physical health.
while among male adolescents, bullying directly related to self-rated While considered a valid measure of adolescents' health, it remains a
health. Interestingly, sex was found to moderate the relationship be- subjective measure. Thirdly, although critical covariates were con-
tween bullying victimization and adolescents' self-rated health, as trolled for, there is always the possibility that significant confounding
bullying was significantly associated with self-rated health for males variables were not included in the model. Fourth, the Fragile Families
but not females after accounting for an individual's mental health and Child Wellbeing Studies over sampled unmarried families at the
status. national level, thus it is not a national representative dataset and its
One possible explanation for this sex-specific pattern is due to the generalizability is limited. Finally, though not the focus of this study,
different types of bullying adolescents face in society given their sex. cyberbullying is another critical social determinant of health which
For male adolescents, the literature indicates they are more likely to be requires further investigation. With consideration of the delineated
victims of physical aggression, while female adolescents are more sus- limitations, this study has several notable strengths. A national sample
ceptible to social and relational exclusions (Stubbs-Richardson, Sinclair, of adolescents was employed rather than being confined to a school-
Goldberg, Ellithorpe, & Amadi, 2018). Therefore, male adolescents based sample which has often been the case in previous studies on
have a greater likelihood of enduring physical injuries caused by bul- bullying victimization. Additionally, adolescent self-rated health was
lying (e.g., fractures, bruises), whereas female adolescents are more used rather than parental rated adolescents' health. This approach
likely to internalize the bullying experience regardless of the lack of better reflects adolescents' own perspective of their health.
physical injury (Swearer, Mebane, & Espelage, 2004). As a result, In summary, findings were in accordance with existing literature
mental health becomes the primary and predominant factor associated and empirical evidence of the relationships between bullying victimi-
with bullying among female adolescents, whereas, in contrast, bullying zation and health. Thus, the study results further strengthen existing
brings both adverse mental and physical impacts to male adolescents. empirical evidence and offer additional contributions to understanding
While the relationship between bullying victimization and mental these relationships with regard to the indirect effects of mental health
health did not significantly vary as a function of sex, a trend toward and moderating effects of sex. Future research should further examine
statistical significance was observed, p = .06. The significant associa- these relationships in various samples within the U.S. as well as in-
tions between bullying victimization and mental health were present vestigate these relationships in the context of cyberbullying.
for both male and female adolescent participants; however, the re-
lationship was stronger for female than male adolescents. This result
offers confirmation to prior research on the importance of

7
A. Zhang, et al. Children and Youth Services Review 104 (2019) 104385

Conflict of interests Forrest, C. B., Bevans, K. B., Riley, A. W., Crespo, R., & Louis, T. A. (2013). Health and
school outcomes during children's transition into adolescence. Journal of Adolescent
Health, 52(2), 186–194.
No conflict of interest. Gini, G., & Pozzoli, T. (2009). Association between bullying and psychosomatic problems:
A meta-analysis. Pediatrics, 123(3), 1059–1065.
COI Statement Gruber, J. E., & Fineran, S. (2008). Comparing the impact of bullying and sexual har-
assment victimization on the mental and physical health of adolescents. Gender Roles,
59(1), 1–13.
All authors claim no conflict of interest. Haberstick, B. C., Young, S. E., Zeiger, J. S., Lessem, J. M., Hewitt, J. K., & Hopfer, C. J.
(2014). Prevalence and correlates of alcohol and cannabis use disorders in the United
States: Results from the national longitudinal study of adolescent health. Drug and
Funding Alcohol Dependence, 136, 158–161.
Halpern, J., Jutte, D., Colby, J., & Boyce, W. T. (2015). Social dominance, school bullying,
No funding to be reported for this study. and child health: What are our ethical obligations to the very young? Pediatrics,
135(S2), S24–S30.
IBM Corp (2016). IBM SPSS statistics for windows, version 24. Armonk, NY: IBM Corp.
Ethics Idler, E. L., & Benyamini, Y. (1997). Self-rated health and mortality: A review of twenty-
seven community studies. Journal of Health and Social Behavior, 38(1), 21–37.
This study is a secondary data analysis thus ethics review does not Jennings, W. G., Song, H., Kim, J., Fenimore, D. M., & Piquero, A. R. (2017). An ex-
amination of bullying and physical health problems in adolescence among south
apply. Korean youth. Journal of Child and Family Studies, 1–12.
Jerdén, L., Burell, G., Stenlund, H., Weinehall, L., & Bergström, E. (2011). Gender dif-
References ferences and predictors of self-rated health development among Swedish adolescents.
Journal of Adolescent Health, 48(2), 143–150.
Jimerson, S. R., & Swearer, S. M. (2009). In D. L. Espelage (Ed.). Handbook of bullying in
Bakker, M. P., Ormel, J., Verhulst, F. C., & Oldehinkel, A. J. (2010). Peer stressors and schools: An international perspective. Routledge.
gender differences in adolescents' mental health: The TRAILS study. Journal of Jylhä, M. (2009). What is self-rated health and why does it predict mortality? Towards a
Adolescent Health, 46(5), 444–450. unified conceptual model. Social Science & Medicine, 69(3), 307–316.
Balázs, J., Miklósi, M., Keresztény, Á., Hoven, C. W., Carli, V., Wasserman, C., ... Cotter, P. Kane, T., & Cantrell, S. (2010). Learning about teaching: Initial findings from the mea-
(2013). Adolescent subthreshold-depression and anxiety: Psychopathology, func- sures of effective teaching project. MET project research paper (pp. 9). Bill & Melinda
tional impairment and increased suicide risk. Journal of Child Psychology and Gates Foundation.
Psychiatry, 54(6), 670–677. Keenan-Miller, D., Hammen, C. L., & Brennan, P. A. (2007). Health outcomes related to
Bauldry, S., Shanahan, M. J., Boardman, J. D., Miech, R. A., & Macmillan, R. (2012). A life early adolescent depression. Journal of Adolescent Health, 41(3), 256–262.
course model of self-rated health through adolescence and young adulthood. Social Kemp, A. H., & Quintana, D. S. (2013). The relationship between mental and physical
Science & Medicine, 75(7), 1311–1320. health: Insights from the study of heart rate variability. International Journal of
Bendheim-Thoman (2018). Data user's guide for the year 15 follow-up wave of the fragile Psychophysiology, 89(3), 288–296.
families and child wellbeing study. Retrieved from https://fragilefamilies.princeton. Kim, Y., Padilla, Y. C., Zhang, A., & Oh, S. (2018). Young children's internalizing and
edu/sites/fragilefamilies/files/ff_public_guide_15.pdf. externalizing behaviors after mothers exit welfare: Comparisons with children of non-
Benyamini, Y. (2011). Why does self-rated health predict mortality? An update on current welfare mothers. Children and Youth Services Review, 86, 316–323.
knowledge and a research agenda for psychologists. Psychology & Health, 26(11), Kumpulainen, K., & Räsänen, E. (2000). Children involved in bullying at elementary
1407–1413. school age: Their psychiatric symptoms and deviance in adolescence: An epidemio-
Boardman, J. D. (2006). Self-rated health among US adolescents. Journal of Adolescent logical sample. Child Abuse & Neglect, 24(12), 1567–1577.
Health, 38(4), 401–408. Little, R. J. (1988). A test of missing completely at random for multivariate data with
Brière, F. N., Pascal, S., Dupéré, V., & Janosz, M. (2013). School environment and ado- missing values. Journal of the American Statistical Association, 83(404), 1198–1202.
lescent depressive symptoms: A multilevel longitudinal study. Pediatrics, 131(3), Løhre, A., Lydersen, S., Paulsen, B., Mæhle, M., & Vatten, L. J. (2011). Peer victimization
702–708. as reported by children, teachers, and parents in relation to children's health symp-
Carr, A. (2015). The handbook of child and adolescent clinical psychology: A contextual ap- toms. BMC Public Health, 11(1), 278.
proach. Routledge. Machmutow, K., Perren, S., Sticca, F., & Alsaker, F. D. (2012). Peer victimisation and
Catalano, R. F., Fagan, A. A., Gavin, L. E., Greenberg, M. T., Irwin, C. E., Jr., Ross, D. A., & depressive symptoms: Can specific coping strategies buffer the negative impact of
Shek, D. T. (2012). Worldwide application of prevention science in adolescent health. cybervictimisation? Emotional and Behavioural Difficulties, 17(3–4), 403–420.
Lancet, 379, 1653–1664. MacKinnon, D. P., Lockwood, C. M., Hoffman, J. M., West, S. G., & Sheets, V. (2002). A
Craig, W., Harel-Fisch, Y., Fogel-Grinvald, H., Dostaler, S., Hetland, J., Simons-Morton, comparison of methods to test mediation and other intervening variable effects.
B., ... Pickett, W. (2009). A cross-national profile of bullying and victimization among Psychological Methods, 7(1), 83.
adolescents in 40 countries. International Journal of Public Health, 54(2), 216–224. McLeod, G. F. H., Horwood, L. J., & Fergusson, D. M. (2016). Adolescent depression, adult
Currie, C., Zanotti, C., Morgan, A., Currie, D., De Looze, M., Roberts, C., ... Barnekow, V. mental health and psychosocial outcomes at 30 and 35 years. Psychological Medicine,
(2009). Social determinants of health and well-being among young people. Health 46(7), 1401–1412.
Behaviour in School-Aged Children (HBSC) Study: International Report from the, 2010, Moore, G. F., Cox, R., Evans, R. E., Hallingberg, B., Hawkins, J., Littlecott, H. J., ...
271. Murphy, S. (2018). School, peer and family relationships and adolescent substance
Danese, A., & Lewis, S. J. (2017). Psychoneuroimmunology of early-life stress: The hidden use, subjective wellbeing and mental health symptoms in Wales: A cross sectional
wounds of childhood trauma? Neuropsychopharmacology, 42(1), 99. study. Child Indicators Research, 11(6), 1951–1965.
Derogatis, L. R., & Savitz, K. L. (1999). The SCL-90-R, Brief Symptom Inventory, and Muthén, L. K., & Muthén, B. O. (2013). Mplus user's guide (6th ed.). (Los Angeles, CA).
Matching Clinical Rating Scales. In M. E. Maruish (Ed.). The use of psychological testing National Center for Education Statistics (2016). Student reports of bullying: Results from
for treatment planning and outcomes assessment (pp. 679–724). Mahwah, NJ, US: the 2015 school crime supplement to the national crime victimization survey.
Lawrence Erlbaum Associates Publishers. Retrieved from https://nces.ed.gov/pubs2017/2017015.pdf.
DiezRoux, A. V. (2012). Conceptual approaches to the study of health disparities. Annual Newman, B. M., & Newman, P. R. (2017). Development through life: A psychosocial ap-
review of public health, 33, 41–58. proach. (Cengage Learning).
Duchesne, A. P., Dion, J., Lalande, D., Bégin, C., Émond, C., Lalande, G., & McDuff, P. Noor-Azniza, I., Malek, T. J., Yahya, S. I., & Farid, T. M. (2011). Moderating effect of
(2017). Body dissatisfaction and psychological distress in adolescents: Is self-esteem a gender and age on the relationship between emotional intelligence with social and
mediator? Journal of Health Psychology, 22(12), 1563–1569. academic adjustment among first year university students. International Journal of
Dukes, R. L., Stein, J. A., & Zane, J. I. (2010). Gender differences in the relative impact of Psychological Studies, 3(1), 78–89.
physical and relational bullying on adolescent injury and weapon carrying. Journal of Oberst, U., Wegmann, E., Stodt, B., Brand, M., & Chamarro, A. (2017). Negative con-
School Psychology, 48(6), 511–532. sequences from heavy social networking in adolescents: The mediating role of fear of
Fairchild, A. J., & MacKinnon, D. P. (2009). A general model for testing mediation and missing out. Journal of Adolescence, 55, 51–60.
moderation effects. Prevention Science: The Official Journal of the Society for Prevention Orpinas, P., McNicholas, C., & Nahapetyan, L. (2015). Gender differences in trajectories
Research, 10(2), 87–99. https://doi.org/10.1007/s11121-008-0109-6. of relational aggression perpetration and victimization from middle to high school.
Fekkes, M., Pijpers, F. I. M., Fredriks, A. M., Vogels, T., & Verloove-Vanhorick, S. P. Aggressive Behavior, 41(5), 401–412.
(2006). Do bullied children get ill, or do ill children get bullied? A prospective cohort Perreira, K. M., Deeb-Sossa, N., Harris, K. M., & Boleen, K. (2005). What are we mea-
study on the relationship between bullying and health-related symptoms. Pediatrics, suring? An evaluation of the CES-D across race/ethnicity and immigrant generation.
117(5), 1568–1574. Social Forces, 83(4), 1567–1601.
Fekkes, M., Pijpers, F. I. M., & Verloove-Vanhorick, S. P. (2004). Bullying behavior and Pouwels, J. L., Scholte, R. H. J., van Noorden, Tirza, H. J., & Cillessen, A. H. N. (2016).
associations with psychosomatic complaints and depression in victims. The Journal of Interpretations of bullying by bullies, victims, and bully-victims in interactions at
Pediatrics, 144(1), 17–22. different levels of abstraction: Interpretations of bullying. Aggressive Behavior, 42(1),
Fong, P., Cruwys, T., Haslam, C., & Haslam, S. A. (2019). Neighbourhood identification 54–65.
and mental health: How social identification moderates the relationship between Radloff, L. S. (1977). The CES-D scale: A self-report depression scale for research in the
socioeconomic disadvantage and health. Journal of Environmental Psychology, 61, general population. Applied psychological measurement, 1(3), 385–401.
101–114. Rambaran, J. A., Hopmeyer, A., Schwartz, D., Steglich, C., Badaly, D., & Veenstra, R.

8
A. Zhang, et al. Children and Youth Services Review 104 (2019) 104385

(2017). Academic functioning and peer influences: A short-term longitudinal study of Moving beyond mean level differences. Bullying in American schools (pp. 37–58).
network–behavior dynamics in middle adolescence. Child Development, 88(2), Routledge.
523–543. Thomas, H. J., Connor, J. P., & Scott, J. G. (2018). Why do children and adolescents bully
Reichman, N. E., Teitler, J. O., Garfinkel, I., & McLanahan, S. S. (2001). Fragile families: their peers? A critical review of key theoretical frameworks. Social Psychiatry and
Sample and design. Children and Youth Services Review, 23(4–5), 303–326. Psychiatric Epidemiology, 53(5), 437–451.
Roland, E. (2002). Bullying, depressive symptoms and suicidal thoughts. Educational Turner, H. A., Finkelhor, D., & Ormrod, R. (2006). The effect of lifetime victimization on
Research, 44(1), 55–67. the mental health of children and adolescents. Social Science & Medicine, 62(1),
Santrock, J. W. (2015). Adolescence (16th ed.). New York, NY: McGraw-Hill. 13–27.
Sawyer, S. M., Afifi, R. A., Bearinger, L. H., Blakemore, S. J., Dick, B., Ezeh, A. C., & Turner, M. G., Exum, M. L., Brame, R., & Holt, T. J. (2013). Bullying victimization and
Patton, G. C. (2012). Adolescence: A foundation for future health. The Lancet, adolescent mental health: General and typological effects across gender. Journal of
379(9826), 1630–1640. Criminal Justice, 41(1), 53–59.
Schmeer, K. K. (2012). Early childhood economic disadvantage and the health of Hispanic US Department of Health and Human Services (2014). Bullying definition. Retrieved from
children. Social Science & Medicine, 75(8), 1523–1530. https://www.stopbullying.gov/what-is-bullying/index.html.
Schofield, T. J., Conger, R. D., Gonzales, J. E., & Merrick, M. T. (2016). Harsh parenting, Veenstra, R., Lindenberg, S., Munniksma, A., & Dijkstra, J. K. (2010). The complex re-
physical health, and the protective role of positive parent-adolescent relationships. lation between bullying, victimization, acceptance, and rejection: Giving special at-
Social Science & Medicine, 157, 18–26. tention to status, affection, and gender differences. Child Development, 81(2),
Shroff, H., & Thompson, J. K. (2006). Peer influences, body-image dissatisfaction, eating 480–486.
dysfunction and self-esteem in adolescent girls. Journal of Health Psychology, 11(4), Verburgh, L., Königs, M., Scherder, E. J., & Oosterlaan, J. (2014). Physical exercise and
533–551. executive functions in preadolescent children, adolescents and young adults: A meta-
Sigurdson, J. F., Undheim, A. M., Wallander, J. L., Lydersen, S., & Sund, A. M. (2015). The analysis. British Journal of Sports Medicine, 48(12), 973–979.
long-term effects of being bullied or a bully in adolescence on externalizing and in- Viner, R. M., Ozer, E. M., Denny, S., Marmot, M., Resnick, M., Fatusi, A., & Currie, C.
ternalizing mental health problems in adulthood. Child and Adolescent Psychiatry and (2012). Adolescence and the social determinants of health. The Lancet, 379(9826),
Mental Health, 9(1), 42. 1641–1652.
Spear, L. P. (2013). Adolescent neurodevelopment. Journal of Adolescent Health, 52(2), Wang, J., Iannotti, R. J., & Nansel, T. R. (2009). School bullying among adolescents in the
7–13. United States: Physical, verbal, relational, and cyber. Journal of Adolescent Health,
Steinberg, L. (2005). Cognitive and affective development in adolescence. Trends in 45(4), 368–375.
Cognitive Sciences, 9(2), 69–74. Williams, L. R., & Anthony, E. K. (2015). A model of positive family and peer relationships
Stubbs-Richardson, M., Sinclair, H. C., Goldberg, R. M., Ellithorpe, C. N., & Amadi, S. C. on adolescent functioning. Journal of Child and Family Studies, 24(3), 658–667.
(2018). Reaching out versus lashing out: Examining gender differences in experiences Zhang, A., De Luca, S., Oh, S., Liu, C., & Song, X. (2019). The moderating effect of gender
with and responses to bullying in high school. American Journal of Criminal Justice, on the relationship between bullying victimization and adolescents' self-rated health:
43(1), 39–66. An exploratory study using the fragile families and wellbeing study. Children and
Suárez, A. L., Feramisco, J. D., Koo, J., & Steinhoff, M. (2012). Psychoneuroimmunology Youth Services Review, 96, 155–162.
of psychological stress and atopic dermatitis: Pathophysiologic and therapeutic up- Zhang, A., Padilla, Y. C., & Kim, Y. (2017). How early do social determinants of health
dates. Acta Dermato-Venereologica, 92(1), 7–18. begin to operate? Results from the fragile families and child wellbeing study. Journal
Swearer, S. M., Mebane, S. E., & Espelage, D. L. (2004). Gender differences in bullying: of Pediatric Nursing, 37, 42–50.