Andersen et al.

BMC Psychology (2015) 3:35

DOI 10.1186/s40359-015-0092-1

RESEARCH ARTICLE Open Access

Bullied at school, bullied at work: a

prospective study
Lars Peter Andersen1*, Merete Labriola2,3, Johan Hviid Andersen1, Thomas Lund1,3 and Claus D. Hansen4

Abstract
Background: The consequences of childhood bullying victimisation are serious. Much previous research on risk factors
for being bullied has used a cross-sectional design, impeding the possibility to draw conclusions on causality, and has
not considered simultaneous effects of multiple risk factors. Paying closer attention to multiple risk factors for being
bullying can provide a basis for designing intervention programmes to prevent or reduce bullying among children
and adolescents.
Methods: Risk factors for bullying were examined by using questionnaire data collected in 2004 and 2007. In 2004,
the participants were aged 14–15 years and 17–18 years in 2007. The baseline questionnaire was answered by 3054
individuals in 2004, and 2181 individuals participated in both rounds. We analysed risk factors for being bullied at the
individual and societal level. Information on the social background of the participants was derived from a national
register at Statistics Denmark.
Results: Several risk factors were identified. Being obese, low self-assessed position in school class, overprotective
parents, low self-esteem, low sense of coherence and low socioeconomic status were risk factors for being bullied at
school. Being overweight, smoking, low self-assessed position in class, low sense of coherence and low socioeconomic
status were risk factors for being bullied at work. However, most associations between risk factors in 2004 and being
bullied in 2007 disappeared after adjustment for being bullied in 2004.
Conclusions: The strongest risk factor for being bullied was being previously bullied. Our results stress the importance
of early prevention of bullying at schools. In addition, attention should be drawn to the role of overprotective parents.
Keywords: Adolescents, Bullying at school, Bullying at work, Risk factors, Prospective study

Background (roughly on a weekly basis) and over a prolonged period

Bullying has been conceptualised as a distinct type of of time (often 6 months) [2]. Given the seriousness of
aggression characterised by a multifaceted form of bullying, prevention of childhood and adolescent bully-
mistreatment seen mostly at schools and at work. The ing has long been considered an important social and
most widely employed definition of bullying emphasises clinical problem.
persistent and repeated negative actions intended to in-
timidate or hurt a weaker person. Bullying includes acts
of deliberate physical aggression (e.g. knocks, punches The prevalence of bullying
and kicks), verbal aggression (e.g. name calling and The prevalence of bullying among adolescents varies
threats), relational aggression (e.g. social isolation and across countries. In two large cross-national studies –
rumour spreading) and cyber aggression (e.g. text messa- The Health Behavior in School-aged Children Survey
ging and e-mailing hurtful messages or pictures) [1]. and the Global School-based Students Health Survey –
The negative interaction must occur relatively often totalling 218,104 students in 66 countries aged between
11 and 15 years, the average prevalence of bullying
* Correspondence: [email protected] victimisation at least once during the past month was
1
Danish Ramazzini Centre, Department of Occupational Medicine, University 32.1 % in the first study, and in second study, 37.4 % of
Research Clinic, Regional Hospital West Jutland, Gl. Landevej 61, 7400
Herning, Denmark
participants reported that they had been bullied at least
Full list of author information is available at the end of the article once within the past 2 months. However, in both studies,
© 2015 Andersen et al. Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Andersen et al. BMC Psychology (2015) 3:35 Page 2 of 15

a large variation in prevalence was found across coun- behaviour, stem from the interplay of a wide range of
tries, from 6 to 41 % in the first study and from 9 to 54 personal, situational and social factors. Therefore, ag-
% in the second [3]. In spite of large differences in the gressive behaviour such as bullying occurs as a result of
prevalence of bullying, the results indicate that too many interactions between the persons involved and factors in
pupils suffer from being victims of aggressive acts the social context that may either facilitate or mitigate
intended to hurt them. the risk of such behaviour [12, 13]. Although a detailed
discussion of the causes of aggression is beyond the
The consequences of being bullied scope of this paper, it is important to underline the
The consequences of childhood bullying victimization complexity of the task of identifying risk factors that
are serious. Both cross-sectional and longitudinal studies are relevant in understanding bullying at schools and
have found that being a victim of bullying is associated workplaces.
with long-term psychological problems, including loneli- Several risk factors for being bullying at schools have
ness, general and social anxiety, diminished self-esteem, been identified (for a review see [1]), but most studies
increased depressive symptoms and more frequent use identifying risk factors for bullying have included only a
of pain medication [4–6]. Finally, being a victim of bully- limited number of risk factors in their statistical models.
ing is an important risk factor for suicidal behaviour in All the risk variables may be correlated with each other,
adolescence and early adulthood [7, 8]. A review based but some may be more important than others in predict-
on 37 studies found that any kind of participation in ing bullying. Therefore, the present study will simultan-
bullying increased the risk of suicidal ideation and/or be- eously examine several different risk factors in young
haviour. The strongest risk was for victim-perpetrators people related to their individual and personal levels, so-
[7]. However, bullies also suffer because severe suicidal cial levels and socioeconomic levels in order to identify
ideation has been found among both those who were the most important risk factors for bullying at schools
bullied and among those who were bullies [9]. and workplaces. This knowledge may be important in
Not surprisingly, many antibullying programmes and order to prevent bullying.
interventions have been implemented in an attempt to Preventing bullying may be good for the bullied as well
reduce the prevalence of being bullied. Unfortunately, the as for the bullies in term of the negative psychological
success of intervention programmes to prevent or mitigate outcomes for both parties.
bullying in childhood and adolescence has been limited. A
synthesis of the existing research on antibullying pro- Risk factors for bullying
grammes concluded that the majority of programmes Potential risk factors for bullying at the individual level
yielded no significant reductions in self-reported bullying, include gender, age, physical appearance and health be-
and therefore only cautious recommendations could be haviour. Regarding gender, the results are inconsistent,
made [9]. and no substantial gender differences have been ob-
served among adolescents in terms of the frequency of
Theoretical frame of reference being bullied either at school or at work [14]. With re-
Given the limited efficacy of bullying intervention pro- spect to age, results from both cross-sectional and longi-
grammes, the purpose of the present study was to more tudinal studies on bullying at schools show that the
closely investigate multiple risk factors for bullying. prevalence of bullying tends to fall with age during ado-
Identifying risk factors can provide a basis for designing lescence [15–17].
intervention programmes to prevent or reduce bullying When teens are asked why some adolescents are
among children and adolescents. bullied, a common response is the deviant appearance
Studies on bullying at schools have identified several of the victim [18]. Overweight and obesity have been
risk factors such as gender, age and deviate appearance found to be associated with an increased risk of being
of the victim; personal characteristics such as low self- bullied in both cross-sectional studies and cohort
esteem and lack of adequate coping skills; social status studies [19–22]. Also, bullying has been found to be
among peers and socioeconomic status in society. How- related to underweight in adolescents [23], and even
ever, bullying is a complex phenomenon, and there is no short pupils are at greater risk of being bullied [24].
single explanation for why some children are bullied by Apparently, any deviance from the physical norm may
others [10]. Furthermore, bullying is conceptualised as a increase the risk of being bullied. Being a smoker also
distinct type of aggressive behaviour, and psychological has been found to increase the risk of being bullied, but
theories of aggression assert that the occurrence of ag- the results of studies on the relation between smoking and
gression can seldom be reduced to one single factor but bullying are inconsistent [25–27].
is more likely to be influenced by several factors simul- At the personal level, self-esteem has been found to be
taneously [11]. Aggressions, like other forms of complex associated with bullying [28]. Self-esteem refers to the
Andersen et al. BMC Psychology (2015) 3:35 Page 3 of 15

global and evaluative view of oneself, and low self- coping skills and lack of social skills needed in order
esteem is associated with a variety of psychological dys- to resolve conflicts in the peer group or with work
functions, whereas high self-esteem is associated with colleagues due to lack of experience with conflict
social adeptness, leadership, higher levels of adjustment resolution in the family. Based on this background,
and good social skills. Therefore, because of poor social we sought to examine whether poor family function-
skills and low levels of adjustment, it seems plausible ing and overprotective parents predict bullying later
that low self-esteem may be a risk factor for being bully- at school or work.
ing. It is, however, unclear whether low self-esteem is a The associations between bullying and coping have
risk factor for bullying or a consequence of being bullied been examined in several studies, and the results show
[29–31]. For instance, among 2326 Italian adolescents, that victims of bullying lack adaptive coping strategies
Brighi et al. [30] found that low global self-esteem was a and more often use avoidant coping or similar strategies
risk factor for victimisation, but on the other hand, an- that might be considered similar [10, 38, 39]. For in-
other study found that victimisation was the most consist- stance, one study found that victims of bullying rarely
ent predictor of low self-esteem [31]. Thus, it is unclear asked for help or talked about what happened, but
whether low self-esteem is a risk factor for or a conse- instead remained passive [40]. Another study found a
quence of being bullied. Furthermore, most studies are positive association between emotionally oriented coping
cross-sectional in design, making causal interpretations strategies and victimisation [10]. Thus, it seems relevant
difficult. In this study, self-esteem was conceptualised as a to study the association between coping and bullying.
risk factor, because it is considered to be a general internal Regarding coping strategies, it is important to include
presentation of social acceptance and rejection and a the victims’ appraisal of the bullying situation because
measure of social functioning. Thus, low self-esteem could appraisals, according to the transactional theory of
be a risk factor for being a target of bullying. stress, determine the coping response, and thus victims’
Studies have demonstrated associations between an in- perceptions of control become important for the imple-
creased risk of being bullied and conflicts with parents mentation of coping strategies [41].
and being from a family characterised by a punitive, The concept of sense of coherence brings the percep-
conflicting and nonsupportive parenting style [10, 32]. tion of control and manageability into consideration.
Additionally, victims’ homes have been found to be char- The concept refers to the individual’s perception of com-
acterised by a higher level of criticism and fewer rules prehensibility, meaningfulness and manageability, the
[33] and having authoritative parents who rarely value last referring to expectations about the availability of
their children and tend not to give them the opportunity adequate resources to cope with stressors. Sense of
to speak up for themselves [34]. Overly protective coherence affects how individuals perceive the events
parenting style could be a risk factor for bullying as well that happen to them, as well as the extent to which they
because parents who are overly protective of their perceive these events as controllable. Persons with a
children and do not let them handle conflicts with peers strong sense of coherence are described as more resist-
by themselves may contribute to the causation of bully- ant to stress and able to cope adaptively [42], and studies
ing [35, 36]. However, the causal direction is unclear have found a direct effect of sense of coherence on stress
because of the cross-sectional design of these studies, [43]. One study found that strong sense of coherence
and protective parenting could also be an outcome of offered protective benefits to targets exposed to bullying
bullying. [44], and another study found that employees with a low
One consequence of inadequate parenting style or sense of coherence more often were subjected to
poor family functioning may be children’s insufficient violence [45]. In the present study, we analysed whether
coping strategies. For instance, a study found a clear re- a low sense of coherence is a risk factor for bullying.
lation between perceived parenting practices and coping In regard to the social context, bullying is related to
in offspring [37]. These researchers found that parenting social status in the group. Indicators of status are social
characterised by warmth and acceptance involved both a preference, popularity, school performance and socio-
high degree of parental monitoring but also parental economic status. Research has found that low levels of
demands for age-appropriate behaviour. Thus, the child social preference and low levels of perceived popularity
may learn that events are to some degree controllable. are associated with an increased risk of being bullied
The result of the study was that the children of parents [46, 47]. However, the cross-sectional nature of the studies
using an accepting and warm parenting style more often cannot exclude the possibility that low social status may be
used problem-focused coping strategies than did children an outcome and not a precursor of bullying, and therefore
who reported that their parents used other rearing styles. more longitudinal studies are needed. Furthermore, a
Therefore, one of the consequences of inadequate parent- meta-analysis concluded that bullied pupils were more
ing style or poor family functioning may be inadequate likely to achieve lower grades than nonbullied pupils. Low
Andersen et al. BMC Psychology (2015) 3:35 Page 4 of 15

grades may reflect interpersonal and social difficulties victim at school [56]. This relationship was seen for
that may increase the risk of bullying. However, the both males and females. Furthermore, another study
cross-sectional nature of the studies cannot exclude found that victims who are bullied at school report
that bullying may lead to mental distress, which could being bullied at work somewhat more often than do
affect school performance [48]. Although the negative others [57]. These results indicate that factors of
relation between victimisation and academic perform- continuity in the risk of victimisation could be more
ance is significant, there are few longitudinal studies related to individual attributes such as low self-
on this topic. Thus, it is unclear whether victimisa- esteem, personality, lack of sufficient coping strat-
tion can be conceptualised as a risk factor or an out- egies and poor ability to establish protective social
come of poor academic performance. In the present relationships than to environmental and organisa-
study, low social status and low school performance tional factors. While both studies were retrospective
were conceptualised as risk factors for bullying; even in design, which could have resulted in recall bias,
though some longitudinal studies have shown that be- the results indicate that being bullied once may be
ing a victim of bullying predicts later academic diffi- an important risk factor for being bullied later. In
culties, there are only limited results supporting this spite of the expansion of research on bullying vic-
notion [49]. timisation, few studies have investigated possible
Finally, the socioeconomic status in society is also re- links between individuals’ experience of previous vic-
lated to bullying, and research has revealed that expos- timisation at school and later victimisation at work
ure to bullying is patterned by socioeconomic status or secondary school [58].
because adolescents from lower socioeconomic status
families are at higher risk of being bullied [50–53] Present study
One explanation could be that inequalities in society Research has revealed several different risk factors for
may lead to more widespread approval of behaviours as- being bullied at school.
sociated with social status differences such as bullying Using data from a prospective cohort study of young
[54]. Furthermore, growing up in a low social status people from the western part of Denmark, this study
family might be associated with more stress in the form examines several risk factors for being bullied at 17–18
of unemployment, divorce, illness and moving, which years of age, making it possible to test the independent
might affect children’s adaptive skills [55], again possibly contribution of each risk factor after adjustment for
increasing the risk of being bullied. covariates in the same domain.
The purpose of the present study was to identify risk
Stability of victimisation from bullying at school and factors measured at age 14–15 for being bullied at age
at work 17–18 at either work or school. Additionally, the study
Given the negative consequences of bullying, it is im- examines the associations between victimisation at the
portant to examine the continuance continued risk of age of 14–15 years and victimisation later at the age of
being bullied. In other words, is being bullied once a risk 17–18 years at a site of higher education or at work.
factor in itself for later victimisation? More specifically, the purposes of the study were as
Relatively little attention has been given to the ef- follows:
fect of being bullied once on later victimisation, and
the few studies that have examined the stability of 1. To examine the prevalence of being bullied at age 15
victimisation from bullying during adolescence have and at age 18 at work and at school
found the risk of being bullied to be relatively stable. 2. To identity the most important risk factors for being
For instance, it was found that being a victim of bullied at age 18, including the following:
bullying at age 8 was associated with victimisation 8 a. Individual risk factors: gender, body mass index,
years later [8, 16]. So far only a few studies have smoking, previously being bullied at age 15
differentiated between victimisation that continues b. Personal risk factors: self-esteem and sense of
from primary school to secondary school and from coherence at age 15
primary school to the workplace. Both at school and c. Social risk factors: parental relations and family
at work, environmental and organisational factors can function at age 15
be sources of bullying, so the stability of victimisation d. Coping strategies: avoidance strategies and
from primary school to the workplace would not support seeking at age 15
necessarily be expected because environmental and e. Indicators of social status: social position in peer
organisational factors change. In spite of this, re- group and social position in society at age 15
searchers have found that the highest risk factor for 3. To examine the continuity of being bullied from
being bullied in the workplace was being a bully age 15 to age 18
Andersen et al. BMC Psychology (2015) 3:35 Page 5 of 15

Methods and had a job, 348 were trainees and 125 had an ordinary
The data used in this study stem from the ongoing West job. Both questions on bullying in 2007 were recorded in
Jutland Cohort Study (Vestliv), a birth cohort study the same way as the 2004 variable.
following a complete regional cohort of adolescents born
in 1989 in the county of Ringkoebing in the western part Health-related lifestyle
of Denmark. The cohort comprised 3681 individuals Two indicators of health-related lifestyle, smoking and
born in 1989, of which 3054 (83 %) answered the base- Body Mass Index were included as possible risk factors
line questionnaire in 2004. Those who had not opted for being bullied. Daily smoking and overweight/obesity
out of the study (N = 3293) were sent the second-round (defined according to the Body Mass Index-for-age
questionnaire in 2007, and 2400 (73 %) answered the values defined by World Health Organisation at age 15)
questionnaire. In all, 2181 individuals participated in were included in the analyses as indicators of unhealthy
both rounds (59 % of the original cohort). These 2181 lifestyle that may be stigmatising and thus lead to a
individuals constitute the primary study base, although higher risk of being bullied.
some of the analyses were carried out on data collected
in only one of the two rounds due to differences in the Indicators of status
available variables. Two traditional measures of parental socioeconomic
The study gathered comprehensive information on the status were derived from information about the partici-
occurrence, severity and impact of manifold symptoms pants’ parents from national registers on income and
of physical and mental health problems, both self- education in 2003 (i.e. the year before baseline data was
reported and register-based [55]. In addition, informa- collected). Household income and highest attained edu-
tion on exposures at school and at home was gathered. cational level in the household were used as measures of
Finally, information on parental socioeconomic status socioeconomic status. If the participants’ parents were
was derived from national registers and linked to the divorced, information on the household in which the
questionnaire data. participants had their place of residence according to
Information on the social background of the partici- the Central Personal Register was used. Information on
pants (e.g. household income, parents’ highest education income was taken from the tax register (recoded into
etc.) was derived from a national register at Statistics tertiles for some of the analyses), and information on
Denmark by using information from the Central Office educational attainment was taken from the Danish
of Civil Registration, in which the respondents are linked Educational Register and recoded into three categories:
to their legal parents or guardians via a personal iden- compulsory school (<10 years of education), high
tification number given to everyone in Denmark at birth school/vocational training (10–12 years) and short,
(or upon entry for immigrants). The study and the linking medium or higher education (>12 years).
of information using the Central Office of Civil Registra- Another mechanism hypothesised to be responsible
tion were approved by the Danish Data Protection Agency for bullying is differences in the social position of ado-
(Study No. 2009-41-3761). The study was approved by the lescents in their peer groups foremost in their school
Danish Data Protection Agency and followed the regula- class. In this study, The MacArthur Scale of Subjective
tions for data storage and protection Social Status – Youth Version was used [59].

Outcome measures Parental relations

Bullying was measured using three single items. In 2004, Parent overprotection was measured using four items
the respondents were asked “How much have you been from the short version of the Parental Bonding Instru-
bullied at school during the last 6 months?” with the ment [60] (Cronbach’s alpha = 0.69). Another aspect of
response categories “Never”, “Once or twice”, “A few parental relations is summed up in the Family Function-
times”, “Once a week” and “Several times a week”. For ing Index, which taps into how well conflicts are re-
use in the analyses, this variable was recorded as “has solved in the family. Family functioning was measured
ever been bullied” and “has not been bullied at all”. In using 12 items from the Family Assessment Device –
2007, two single measures were used with slightly differ- General Functioning scale (FAD-GF) [61]. (Cronbach’s
ent wording. The respondents were asked: “How much alpha = 0.86). Both indicators of parental relations were
have you been bullied in an unpleasant way at school dichotomised for the multivariate analyses using the
during the last 6 months?” with the same response median as cut-off point.
categories as in 2004. In addition, the question was
asked about bullying at work for the respondents who Personal psychological characteristics
had a job at the time of the baseline questionnaire. 478 Two different aspects of the adolescents’ personal
of those studied did not have a job, 1305 were at school characteristics were included in the analyses. First of all,
Andersen et al. BMC Psychology (2015) 3:35 Page 6 of 15

self-esteem – measured by the 6-item version of Rosen- consistency was a little low, but our sample’s consistency
berg’s global self-esteem scale [62] – was included appeared to be higher than that of other translations.
(Cronbach’s alpha = 0.82). Secondly, the four items tap- Cronbach’s alphas for the coping subscales were similar to
ping into the meaningfulness dimension of Antonovsky’s the alpha values reported by Carver, supporting the reli-
construct SOS were also included [42] (Cronbach’s ability of our measures despite the somewhat low alpha
alpha = 0.62). One could argue that the measure of in- coefficients. Both scales were created using the mean of
ternal consistency was a little low. However, compar- the items, thus yielding two scales with scores between 1
ing our results with those of other translations, the and 4, with higher scores indicating higher levels of that
consistency actually appeared to be somewhat higher type of coping. All four measures of psychological char-
in our sample [63]. Coping was measured using two acteristics were dichotomised for use in the multivari-
scales based on the Brief COPE Scale used in a previ- ate analyses using the median as cut-off value.
ous paper based on data from this cohort [55]. The ori-
ginal subscales were divided into two coping dimensions Statistical analyses
that emphasised either an “active” approach to problem Table 1 shows all the characteristics of the study sample.
solving, generally considered to be more adaptive, or Multiple logistic regression models were used to study
“avoidance”-based approach, considered to be less adap- the association between bullying and the two dependent
tive. The six items from the subscales “active coping”, variables. For use in the logistic regression models, all of
“planning” and “positive reframing” were combined to the scales were dichotomised using the median score as
form the “active” coping scale (Cronbach’s alpha = 0.75), the cut-off point, e.g. creating a variable coded 1 if the
and the four items from the subscales “self-distraction” respondent had levels of self-esteem below the median.
and “behavioural disengagement” were used to form the Four models were tested. The bivariate association was
“avoidance” coping scale (Cronbach’s alpha = 0.53). Once calculated between each of the independent variables
again, one could argue that the measure of internal and bullying (Model 1). After this procedure, the
Table 1 Characteristics of study population of children at age 15 years. N = 2,278
Variable name Categories N Pct/mean (SD)
Gender Female 1,237 54.3
Male 1,041 45.7
Lifestyle factors
Body Mass Index for age Severe thinness 57 2.9
Normal weight 1,640 83.0
Overweight 216 10.9
Obese 62 3.1
Smoking habits Not smoker smoke 1,857 89.3
Daily smoker 222 10.7
Indicators of status
School performance Mean (SD). scale; 0-13 8.99 1.17
Self-assessed position in school class Mean (SD). scale;1-10 7.08 1.72
Parental educational level <10 year 255 11.2
10-12 years 1,169 51.4
>13 852 37.4
Household income, DKR Mean (SD) 2,276 573,574 (250,511)
Social relations with parents
Overprotective parents Mean (SD) scale; 4-16 8.10 2.63
Family functioning Mean (SD), scale;1-4 1.72 0.50
Personal psychological characteristics
Self-esteem Mean (SD) scale: 6-24 19.08 2.90
Sense of Coherence (Meaningfulness) Mean (SD) scale: 5-20 14.37 2.14
Active coping Mean (SD) scale: 1-4 2.65 0.52
Avoidance coping Mean (SD) scale: 1–3.5 1.94 0.47
Andersen et al. BMC Psychology (2015) 3:35 Page 7 of 15

Table 2 Prevalence of bullying at school and at work at ages 15 and 18

Variable Level N Pct
Bullied at school during last 6 months (age 14–15) Not bullied 2,255 74.6
Bullied 766 25.4
Frequently bullieda 76 9.9
Bullied in an unpleasant way at school during last 6 months (age 17–18) Not bullied 1,843 87.8
Bullied 256 12.2
Frequently bullieda 12 4.7
Bullied in an unpleasant way at work during last 6 months (age 17–18) Not bullied 1,535 91.4
Bullied 144 8.6
Frequently bullieda 8 5.5
a
Frequently bullied = Bullied at least once a week/Ever bullied

independent variables were entered into models by the shown. The risk of experiencing bullying at school at
abovementioned themes to test the independent contri- age 17–18 was approximately twice as high for those
bution of each of the variables after taking into account who were bullied at school at age 14–15. The associ-
other aspects of the same processes (Model 2). Bullying ation was somewhat stronger for bullying at work.
experience in 2004 was entered into the domain-specific Nearly 45 % of those who experienced bullying at
models (Model 3). Last, a final model in which step- school at age 14–15 also experienced bullying in their
wise forward selection was used to reduce the num- higher educational track at age 17–18 or at work 3
ber of independent variables was created (Model 4). years later.
The Hosmer-Lemeshow goodness-of-fit was used to Table 4 displays risk factors for bullying experiences at
evaluate the quality of the models [64]. All in all, the school at age 17–18. Only three variables remained sta-
sample used for the multivariate analyses consisted of tistically significantly associated with bullying after mu-
1853 and 1376 respondents being bullied at school tually adjusting for each other and for being bullied at
and at work, respectively; participants were excluded age 14–15. First of all, having been bullied at school at
on a model-based basis, so the number of participants in age 14–15 raises the risk of experiencing being bullying
each model varied according to the number of missing 3 years later by factor of 3 even when after taking other
data elements on any of the variables. The analyses were possible factors into account. Secondly, having a parent
performed using STATA 13 (Stata Statistical Software: with a parenting style more overprotective than the
Release 13. College Station, TX, USA: StataCorp LP). median also raised the risk of being bullied. For those
scoring in the highest decile on the scale, the risk of be-
Results ing bullied was more than twice as high as it is for those
Table 2 shows the prevalence of being bullied at age having the least overprotective parents. Many of the
14–15 at school and age 17–18 at school and at work. included variables were bivariately associated with bully-
Due to changes in the wording of the question, it was not ing, e.g. Body Mass Index and measures of social pos-
possible to compare the prevalence across years. At age ition in the school class attended at age 14–15. These
17–18, it appeared that bullying was more prevalent in the associations were diluted and largely disappeared after
school context than at the adolescents’ workplaces. A very mutually adjusting for all the variables used, as well as
low number of individuals were exposed to weekly after entering the variable indicating whether the re-
bullying behaviour; less than 10 % of those reporting spondents had experienced bullying at age 14–15.
any bullying were bullied weekly or more often. Finally, Table 5 shows the associations between the
In Table 3, the association between bullying at age independent variables and experiencing bullying at the
14–15 and bullying in either setting at age 17–18 is workplace at age 17–18. The strongest association

Table 3 Association between bullying at age 15 and bullying at age 18. Percentages and prevalence proportion ratios (PPRs) with
95 % confidence intervals
Proportion bullied at school (age 18) Proportion bullied at work (age 18)
N (Pct) PPR (95 % CI) N/Pct PPR
Not bullied at school age 15 69 (6.0 %) 1 (ref) 123 (8.5 %) 1 (ref)
Bullied at school at age 15 56 (15.1 %) 1.99 (1.56-2.54) 107 (23.4 %) 2.23 (1.86-2.67)
Andersen et al. BMC Psychology (2015) 3:35 Page 8 of 15

Table 4 Associations between and lifestyle, personal characteristics, peer group and parental relations at age 15 and being bullied
at school age 18. Odds ratios (OR) obtained by logistic regression with 95 % confidence intervals
Model 1 Model 2a Adjusted Model 3a Adjusted for Model 4** Forward
Bivariate for covariates in covariates in same stepwise selection
same domain domain and bullying
OR (95% CI) OR (95 % CI) OR (95 % CI) OR (95 % CI)
Gender
Girls Ref 1.00 1.00 1.00
Boys 1.36 (1.04-1.76) 1.37 (1.04-1.82) 1.40 (1.04-1.89)
Lifestyle factors
Body Mass Index
Underweight 1.18 (0.53-2.66) 1.18 (0.53-2.65) 1.17 (0.51-2.68) removed
Normal weight Ref 1.00 1.00 1.00
Overweight 1.35 (0.88-2.06) 1.35 (0.89-2.07) 1.22 (0.79-1.89)
Obesity 2.27 (1.17-4.41) 2.37 (1.17-4.40) 1.68 (0.85-3.33)
Smoking
Does not smoke Ref 1.00 1.00 1.00 removed
Smoke 1.14 (0.73-1.77) 0.98 (0.46-2.08) 0.95 (0.44-2.06)
Social position in peer group
School performance
School performance above median Ref. 1.00 1.00 1.00 removed
School performance below median 1.26 (0.95-1.67) 1.06 (0.78-1.43) 1.04 (0.76-1.42)
Self-assessed position in schoolat school class removed
Self-assessed position in school at school Ref. 1.00 1.00 1.00
class above median
Self-assessed position in school at school 1.49 (1.12-1.99) 1.46 (1.08-1.98) 1.18 (0.86-1.62)
class below median
Parental relations
Over-protection
Parents less over-protective than median Ref. 1.00 1.00 1.00 1.00
Parents more over-protective than median 1.70 (1.29-2.25) 1.52 (1.12-2.06) 1.41 (1.03-1.92) 1.46 (1.08-1.95)
Family Functioning
Family functioning above median Ref. 1.00 1.00 1.00 removed
Family Functioning below median 1.51 (1.13-2.01) 1.25 (0.92-1.70) 1.14 (0.83-1.56)
Personal psychological characteristics
Self-esteem
Self-esteem above median Ref. 1.00 1.00 1.00 1.00
Self-esteem below median 1.66 (1.23-2.22) 1.51 (1.10-2.06) 1.33 (0.97-1.84) 1.41 (1.03-1.94)
Sense of coherence (meaningfulness)
Sense of coherence above median Ref. 1.00 1.00 1.00 removed
Sense of coherence below median 1.55 (1.17-2.05) 1.38 (1.02-1.87) 1.25 (0.92-1.70)
Active coping
Active coping above median 1.00 1.00 1.00 removed
Active coping below median 1.17 (0.88-1.55) 0.95 (0.70-1.29) 0.94 (0.69-1.28)
Avoidance coping
Avoidance coping below median Ref. 1.00 1.00 1.00 removed
Avoidance coping above median 1.39 (1.04-1.84) 1.27 (0.95-1.71) 1.20 (0.89-1.63)
Socioeconomic status
Andersen et al. BMC Psychology (2015) 3:35 Page 9 of 15

Table 4 Associations between and lifestyle, personal characteristics, peer group and parental relations at age 15 and being bullied
at school age 18. Odds ratios (OR) obtained by logistic regression with 95 % confidence intervals (Continued)
Household income 2003 removed
Lowest tertile 1.45 (1.03-2.03) 1.32 (0.92-1.89) 1.18 (0.80-1.74)
Middle tertile 1.40 (1.03-1.90) 1.38 (1.01-1.90) 1.25 (0.90-1.76)
Highest tertile Ref 1.00 1.00 1.00
Household highest attained removed
education 2003
<10 year 1.65 (1.09-2.49) 1.47 (0.94-2.29) 1.50 (0.92-2.43)
10-12 years 1.03 (0.77-1.37) 0.95 (0.71-1.28) 0.93 (0.68-1.27)
>13 years Ref 1.00 1.00 1.00
Being bullied at age 14–15
Not being bullied at age 14–15 Ref 1.00 1.00
Being bullied at age 14–15 3.14 (2.34-4.23)
Hosmer-Lemeshow Goodness-of-fit Chi-square 23.44 = 0.02)
2
Nagelkerke Pseudo R 0.06
a
Adjusted for other factors in the same domain
**Adjusted for all variables with a p-value < 0.20

observed for bullying at work was gender; boys had a with 4.7 % being frequently bullied at school and 5.5 %
more than twofold higher risk of experiencing bullying being bullied at work. This is also in accordance with pre-
in the adjusted analysis. Having experienced bullying at vious studies, which have documented a decrease in the
school at age 14–15 raised the risk of being bullied at frequency of bullying during school life [15, 17]. Further-
work, and this was also seen for bullying at secondary more, at age 17–18 bullying was more frequent at work
school. The association, however, was somewhat lower than at school.
than the association between bullying at primary and The wording of the questions in 2007 was somewhat
secondary school. Parental relations also predict experi- stricter, which meant that it was not possible to directly
encing bullying: having more troublesome relationships compare the prevalence of bullying at school between
with parents characterised by conflict and lack of com- the two rounds.
munication raised the risk of bullying at work. And fi- The second purpose of the study was to identify the
nally, scoring below the median on the meaningfulness most important risk factors for being bullied at age 18,
dimension of the sense of coherence scale increased including risk factors at several levels. As mentioned
the risk of being bullied at work even after taking earlier, aggressive behaviour such as bullying stems
into account the experience of bullying at age 14–15. from the interplay of a wide range of personal, situ-
As was the case with the risk factors for bullying in ational and social factors. In model 2, in which we
school, several variables were bivariately associated mutually adjusted for other risk factors, we found
with experiencing bullying at work at age 17–18. The that obesity, low self-assessed position in school class,
effects of Body Mass Index, daily smoking and overprotective parents, low self-esteem, low sense of
measures of social position in the peer group were all coherence and middle socioeconomic status were all
diluted after mutually adjusted for each other and for significant risk factors for being bullied 3 years later
the experience of bullying at age 14–15. There was at school. In this same model (model 2), nearly iden-
still, however, a slightly elevated risk for obese adoles- tical risk factors (plus being a smoker) were identified
cents to experience bullying at work, even if the at work. The results underline that bullying acts
estimate was somewhat fragile. occur as a result of complex interactions between the
persons involved and factors in the social context [12,
Discussion 13], and furthermore, the results indicate somewhat
The first purpose of the present study was to examine the similar risk factors for bullying across organisational
prevalence of being bullied at age 14–15 and age 17–18 at settings (i.e. school and workplace).
work and school. We found that that nearly 10 % of the In the next section, we will discuss the most important
participants reported being frequently bullied at age risk factors.
14–15 during the last 6 months. This is in line with We found that physical appearance was a risk factor
other studies. At age 17–18, the prevalence had decreased, for being bullied both in school and at work. Obesity
Andersen et al. BMC Psychology (2015) 3:35 Page 10 of 15

Table 5 Associations between and lifestyle, personal characteristics, peer group and parental relations at age 15 and being bullied
at work age 18. Odds ratios (OR) obtained by logistic regression with 95 % confidence intervals
Model 1 Bivariate Model 2aAdjusted Model 3aAdjusted Model 4** Forward
for covariates in for covariates in same stepwise selection
same domain domain and bullying
OR (95 % CI) OR (95 % CI) OR (95 % CI) OR (95 % CI)
Gender
Girls Ref 1.00 1.00 1.00
Boys 2.39 (1.68-3.41) 2.11 (1.44-3.09) 2.23 (1.48-3.35)
Lifestyle factors
Body Mass Index
Underweight 1.05 (0.32-3.47) 1.07 (0.32-3.55) 0.95 (0.28-3.23) 0.84 (0.24-2.95)
Normal weight Ref 1.00 1.00 1.00 1.00
Overweight 1.95 (1.17-3.23) 1.94 (1.17-3.23) 1.84 (1.10-3.09) 1.77 (1.04-3.02)
Obesity 2.04 (0.84-4.96) 2.14 (0.88-5.22) 1.76 (0.71-4.35) 1.43 (0.56-3.64)
Smoking
Does not smoke Ref 1.00 1.00 1.00 1.00
Smoke 1.98 (1.02-3.85) 2.16 (1.10-4.24) 2.23 (1.12-4.43) 1.97 (0.94-4.11)
Social position in peer group
School performance
School performance above median Ref. 1.00 1.00 1.00 removed
School performance below median 1.33 (0.91-1.94) 1.11 (0.74-1.68) 1.08 (0.71-1.64)
Self-assessed position in school class removed
Self-assessed position in school class Ref. 1.00 1.00 1.00
above median
Self-assessed position in school class 1.84 (1.24-2.73) 1.78 (1.18-2.68) 1.49 (0.98-2.26)
below median
Parental relations
Over-protection
Parents less over-protective than median Ref. 1.00 1.00 1.00 removed
Parents more over-protective than median 1.40 (0.97-2.02) 1.11 (0.75-1.64) 1.03 (0.69-1.54)
Family Functioning
Family functioning above median Ref. 1.00 1.00 1.00 removed
Family Functioning below median 2.03 (1.37-3.01) 1.95 (1.29-2.95) 1.87 (1.23-2.85) 1.69 (1.10-2.59)
Personal psychological characteristics
Self-esteem
Self-esteem above median Ref. 1.00 1.00 1.00 1.00
Self-esteem below median 1.65 (1.11-2.45) 1.37 (0.90-2.10) 1.29 (0.84-2.00) 1.60 (1.03-2.49)
Sense of coherence (meaningfulness)
Sense of coherence above median Ref. 1.00 1.00 1.00 removed
Sense of coherence below median 1.99 (1.36-2.91) 1.70 (1.13-2.57) 1.56 (1.03-2.37)
Active coping
Active coping above median Ref. 1.00 1.00 1.00 removed
Active coping below median 1.40 (0.96-2.05) 1.13 (0.75-1.71) 1.09 (0.72-1.64)
Avoidance coping
Avoidance coping below median Ref. 1.00 1.00 1.00 removed
Avoidance coping above median 1.32 (0.90-1.92) 1.17 (0.79-1.74) 1.14 (0.76-1.69)
Socioeconomic status
Andersen et al. BMC Psychology (2015) 3:35 Page 11 of 15

Table 5 Associations between and lifestyle, personal characteristics, peer group and parental relations at age 15 and being bullied
at work age 18. Odds ratios (OR) obtained by logistic regression with 95 % confidence intervals (Continued)
Household income 2003
Lowest tertile 2.06 (1.29-3.28) 1.89 (1.15-3.11) 1.64 (0.96-2.80) 1.54 (0.89-2.67)
Middle tertile 2.17 (1.41-3.33) 2.16 (1.39-3.36) 1.95 (1.22-3.10) 2.10 (1.30-3.39)
Highest tertile Ref. 1.00 1.00 1.00 1.00
Household highest attained removed
education 2003
<10 year 1.83 (1.07-3.11) 1.44 (0.81-2.55) 1.39 (0.73-2.65)
10-12 years 1.13 (0.77-1.66) 0.94 (0.63-1.40) 1.01 (0.66-1.55)
>13 years Ref. 1.00 1.00 1.00
Being bullied at age 14–15
Not being bullied at age 14–15 1.00
Being bullied at age 14–15 2.09 (1.39-3.14)
Hosmer-Lemeshow Goodness-of-fit Chi-square 166.26 = 0.48)
2
Nagelkerke Pseudo R 0.09
a
Adjusted for other factors in the same domain
** Adjusted for all variables with a p-value < 0.20

was a risk factor for being bullied at school, and over- One explanation might be that being a smoker is associ-
weight was a risk factor for being bullied at work. ated with lower social and economic status [68], which is
After controlling for previously being bullied, the often found to be a risk factor for being bullied [47, 69].
association between obesity and bullying at school Another plausible interpretation is that smoking is in-
disappeared, but overweight remained a risk factor for creasingly being frowned upon because legislation regulat-
being bullied at work. Several studies have found that ing smoking at workplaces was tightened in 2007 when
overweight is a risk factor for being bullied [20–22]. the adolescents were 17–18 years old. This would lead to
The possible link between overweight and bullying smoking behaviour being more and more at odds with the
could be that overweight adolescents have poorer psy- prevailing norms in society.
chological well-being and more depressive feelings, Social position in the peer group was found to be a
which would put them at increased risk of being risk factor for being bullied both at work and at school.
bullied and furthermore affect their perceptions of Previous studies have demonstrated that there is an as-
other people’s teasing [65, 66]. Thus, it could be that sociation between low social status and increased risk of
it is the psychological consequences of being over- bullying [47, 69], and the present study adds to the
weight, not simply being overweight, that increase the current knowledge that low social position is a signifi-
risk of being bullied. cant risk factor for being bullied irrespective of whether
At work, overweight remained a risk factor for being the adolescents at age 17–18 are at school or at work.
bullied even after adjusting for covariates such as self- One explanation could be that perceived popularity re-
esteem and sense of coherence. At work at least, it is flects dominance, and therefore it may be easy for a
more likely to be the physical appearance of overweight popular child or adolescent to bully others who are low
that is the determining factor behind the observed asso- in popularity without fear of being sanctioned by peers
ciation. One mechanism might be that overweight [46]. Consequently, bullying could be one way to main-
workers may have difficulties keeping up with the rapid tain high status. Most studies examining the importance
tempo of unskilled piecework jobs and thus be at risk of of social position in peer groups in relation to bullying
being bullied. For instance, in unskilled piecework in the have used cross-sectional designs [46, 47], but this study
construction sector, if a crew member has difficulties adds to the existing literature by using a longitudinal de-
keeping up the tempo, he or she is at risk of informal sign, making the results more robust. However, after ad-
sanctions [67]. justment for previously being bullied, the associations
The results show that being a smoker at age 15 is a were weaker. Furthermore, low school performance was
significant risk factor for being bullied at age 18, but not shown to be a risk factor for bullying in our study.
only at work. This association remains significant Social position in society, measured as socioeco-
even after controlling for being bullied at age 14–15. nomic status, was also found to be a risk factor both
Andersen et al. BMC Psychology (2015) 3:35 Page 12 of 15

at work and at school. The negative association was bullying remained moderately strong or strong in
seen for both low income and low education among most cases.
parents, but it was diluted after adjusting for previ- The third purpose of the study was to examine the
ously being bullied. continuity of being bullied from age 15 to age 18. The
Having parents with a parenting style classified as results showed that the highest risk (besides gender) for
overprotective was found to be a risk factor for being being bullied at age 17–18 at secondary school or at
bullied at school, and low family function was a risk work was associated with being bullied at age 15. Appar-
factor for being bullied at work even after adjusting ently, being bullied once significantly increased the risk
for being bullied at age 14–15. In other words, par- for future exposure to bullying. This is similar to previ-
enting style at age 14–15 increases the risk of being ous findings demonstrating that being a victim of bully-
bullied 3 years later. Other studies have also demon- ing increases the risk of later victimisation [8, 16, 56],
strated an association between parenting style and which suggests a continuity of being bullied. In this
being bullied [10, 32], and parental overcontrol has study, the highest risk factor for being bullied was previ-
been found to be a risk factor for bullying among ously being bullied, and the results remained stable after
adolescent girls. The mechanism could be that paren- adjustment for other risk factors. This degree of individual
tal overcontrol and protection limits adolescents’ op- consistency of being bullied in different environments and
portunities to interact with certain kinds of peers and at different ages points to factors of continuity in the
prevents adolescents from developing a diverse set of risk of victimisation. One explanation may be that
social skills [70]. Another mechanism could be that early bullying affects the perception of the self and
children, through their experiences at home (e.g. ob- relations to others, or the answer may be in individ-
servational learning [71]), learn and reproduce insuffi- ual attributes such as temperament, self-esteem, the
cient ways of solving conflicts, which they reproduce ability to form protective relationships and coping.
at school or at work with their peers. Thus, parenting Temperament is a stable characteristic of the individ-
style may either increase the risk of being bullied or ual [72], but coping and self-esteem also seem to be
protect children from being bullied. rather stable over time. The stability of the coping
Low self-esteem was found to be a risk factor for being response in adolescence was studied by Kirchner et
bullied both at school and at work. This is in line with al. [73], who found that coping responses were quite
previous findings [29, 30]. In the present study, the asso- stable over time, especially avoidance coping, which
ciation was weakened after adjusting for previously being may explain the continuing risk of bullying. Similar
bullied. The mechanism could be that being bullied results have been found by other researchers [74, 75].
damages self-esteem, which would make it harder to Furthermore, Alsaker and Olweus [76] found that ad-
manage teasing and bullying from peers. olescents’ negative self-evaluation (self-esteem) is
Low levels of a sense of coherence were also a risk likely to become relatively more crystallised with in-
factor for being bullied both at school and at work. creasing age. These findings suggest the importance
People with a strong sense of coherence are more resist- of certain individual characteristics as risk factors for
ant to stress and may be more insensitive to stressors later bullying at school and at the workplace.
like bullying [42]. We found that a weak sense of coher- However, there is also a substantial degree of discon-
ence increased the risk of being bullied. This result is in tinuity in being bullied. Many victims of early school
accordance with a Danish study that found that bullying do not become workplace victims or victims of
employees subjected to violence had a weaker sense of bullying at secondary school. This suggests that although
coherence [45]. stable individual factors may have some importance in
Based on the size of the odds ratio, the increased explaining victimisation risk, contextual factors such as
risk for being bullied seems to be embedded first and immediate environment and social support are also im-
foremost in individual and personal characteristics portant. This is supported by research that has shown
(obesity, low self-esteem, overprotective parents, low that the most important risk factors for the development
sense of coherence) and to a lesser degree in social of bullying in the workplace on the team and organisa-
contexts (low self-assessed position, low/middle socio- tional level were leadership style, norms and values, and
economic status). Most likely the risk factors for be- communication and social climate [77]. Furthermore,
ing bullied reinforce each other, and future research employees that experienced bullying had lower percep-
may address the interaction between risk factors. tions of their work environment in general, especially in
After adjusting for being bullied at baseline, most of relation to trust, cooperation, conflict resolution and
the aforementioned associations became insignificant, justice in the organisation [78]. Thus, a great deal of the
but irrespective of statistical significance, the size of variance in workplace bullying may be due to environ-
the associations (odd ratio) between risk factors and mental factors.
Andersen et al. BMC Psychology (2015) 3:35 Page 13 of 15

Limitations Finally, the discontinuity in being bullied may be a

Although the study reveals significant associations be- focus for future research because many victims of early
tween risk factors and bullying, caution about causal school bullying do not become workplace victims or
inference is warranted. Firstly, it can be argued that a victims of bullying at secondary school. It is importance
3-year gap between exposure to bullying and the measure- to identify protective factors at both the personal and
ment of bullying is too long. However, a certain amount organisational levels.
of time between baseline and follow-up data is necessary Another important addition to our paper would have
as one main criterion of the definition of being bullied is been the implementation of more sophisticated statis-
the prolonged nature of the negative experience [2], and tical modelling, such as structural equation modelling,
negative acts often develop over a long time span [79]. In e.g. using cross-lagged models. This would have made it
addition, it is unknown to what extent the associations re- possible to gain more knowledge about the associations
ported in this paper are weakened due to dropouts among between the risk factors examined in this paper and the
those being bullied at age 14–15 from the first to the two outcomes: bullying at work and bullying at school.
second questionnaires, but the response rate was still 59 % This type of analysis has the potential to better illumin-
of the original cohort. ate the possible causal associations between the risk
Furthermore, the results should be interpreted with factors for bullying identified in this study. However, we
some caution because the study is based only on self- were not able to use cross-lagged modelling of bullying
reported measures, which increases the risk of common for two reasons. First of all, in the first round of the
methods bias [80]. Furthermore, it is also worth noting questionnaire there was only one question on bullying
that the measurement of bullying consisted of a single (because the adolescents were too young (14–15) to
item, leaving it up the individual participant to define have a job in which they spent enough time to allow for
the concept of bullying. However, self-reported measures an exposure to bullying). Secondly, the wording of the
are common in this area of research and seem to yield questions was slightly different, which also prohibit a
reliable and valid results [81]. proper cross-lagged analysis. Finally, if the aim was to
Finally, the data did not contain information on per- cross-lag socioeconomic status and bullying, this would
sonality traits such as neuroticism that could affect the also be difficult because the measure of socioeconomic
reporting of bullying and the predictor variable. status is of the educational level and income in the
household in which the adolescents are living. This
means that those adolescents that have moved out of
Future research their parent’s homes would no longer have a comparable
The results stress the importance of more prospective socioeconomic status measure, which again makes the
cohort studies. The underlying and longitudinal psy- cross-lagged approach very difficult because of the
chosocial mechanisms of being bullied are unclear transitional phase the adolescents were in at the time
and may be mediated through a decreased ability to of participation in the study.
adapt sufficiently. More prospective studies with sev-
eral time lags are needed to determine the modera-
ting and mediating psychological processes between Conclusion
risk factors and bullying and to determine whether The results show that the most important risk factors for
the increased risk of early victimisation from being being bullied at age 17–18, whether at work or in school,
bullied remains a risk factor into adulthood. Further- were being a boy and previously being bullied. This result
more, future research must focus on factors related to stresses the importance of early prevention of bullying at
the individual or family context that can moderate schools. In addition, having overprotective parents was a
negative factors and make the individual more resi- risk factor for being bullied at school. Being overweight,
lient. Finally, future research must examine the im- smoking, experiencing low family function and having low
portance of personality. So far the results concerning socioeconomic status were additional risk factors for being
personality as a risk factor are inconsistent, and there is a bullied at work.
need for more prospective studies to determine whether In summary, the increased risk for being bullied seems
neuroticism predisposes children to victimisation in to be embedded first and foremost in individual and per-
school or whether victimisation affects the develop- sonal characteristics, to a lesser degree in the social con-
ment of a vulnerable personality [82]. In addition, text, but most importantly, in the experience of previously
there is a need for qualitative studies exploring in being bullied.
detail the psychological processes leading from being
bullied once to an increased vulnerability to later Competing interests
victimisation. No competing interests or financial competing interests are declared.
Andersen et al. BMC Psychology (2015) 3:35 Page 14 of 15

Authors’ contributions 15. Analitis F, Velderman MK, Ravens-Sieberer U, Detmar S, Erhart M, Herdman
LPA interpreted data, drafted the manuscript and made the final version M, et al. Being Bullied: Associated Factors in Children and Adolescents 8 to
ready for publication. JHA participated in the conception and design of the 18 Years Old in 11 European Countries. Pediatrics. 2009;123(2):569–77.
study, interpreted the data, revised the draft for important content and 16. Sourander A, Jensen P, Roenning JA, Niemela S, Helenius H, Sillanmaki L, et
made the final version for publication. CDH participated in the conception al. What Is the Early Adulthood Outcome of Boys Who Bully or Are Bullied
and design of the study, performed the statistical analyses, interpreted the in Childhood? The Finnish “From a Boy to a Man” Study. Pediatrics.
data, drafted the section on methods and results, revised the draft for 2007;120(2):397–404.
important content and made the final version for publication. TL and ML 17. Pellegrini AD, Long JD. A longitudinal study of bullying, dominance, and
interpreted data, revised the draft for important content and made the final victimization during the transition from primary school through secondary
version for publication. All agree to be accountable for all aspects of the school. Br J Dev Psychol. 2002;20(2):259–80.
work. All authors read and approved the final manuscript. 18. Frisen A, Jonsson A, Persson C. Adolescents’ perception of bullying: Who is
the victim? Who is the bully? What can be done to stop bullying?
Adolescence. 2007;42(168):749–61.
Acknowledgements 19. Juvonen J, Wang Y, Espinoza G. Bullying experiences and compromised
This research was supported by The Danish Working Environment Research academic performance across middle school grades. J Early Adolesc.
Fund. Grant number 5253. 2011;31(1):152–73.
20. Janssen I, Craig WM, Boyce WF, Pickett W. Associations between overweight
Author details and obesity with bullying behaviors in school-aged children. Pediatrics.
1 2004;113(5):1187–94.
Danish Ramazzini Centre, Department of Occupational Medicine, University
Research Clinic, Regional Hospital West Jutland, Gl. Landevej 61, 7400 21. Griffiths LJ, Wolke D, Page AS, Horwood JP. Obesity and bullying: different
Herning, Denmark. 2Aarhus University, Department of Public Health, effects for boys and girls. Arch Dis Child. 2006;91(2):121–5.
Universitets Parken, Aarhus 8000, Denmark. 3CFK Public Health and Quality 22. GUO QZ, MA WJ, NIE SP, XU YJ, XU HF, ZHANG YR. Relationships between
Improvement Central Denmark Region, P. P. Orumsgade 11, Aarhus 8000, Weight Status and Bullying Victimization among School-aged Adolescents
Denmark. 4Department of Sociology & Social Work, Kroghstrædet 5, Aalborg in Guangdong Province of China. Biomed Environ Sci. 2010;23(2):108–12.
University, Aalborg 9220, Denmark. 23. Wang J, Iannotti RJ, Luk JW. Bullying Victimization Among Underweight and
Overweight U.S. Youth: Differential Associations for Boys and Girls. J Adolesc
Received: 16 November 2014 Accepted: 30 September 2015 Health. 2010;47(1):99–101.
24. Voss LD, Jean M. Bullying in school: are short pupils at risk? Questionnaire
study in a cohort. BMJ. 2000;320.
25. Roberto F, Lyndall M, Chris R, Adrian B. Bullying behaviour and psychosocial
References health among school students in New South Wales, Australia: cross
1. Cook CR, Williams KR, Guerra NG, Kim TE, Sadek S. Predictors of bullying and sectional survey. BMJ. 1999;319.
victimization in childhood and adolescence: A meta-analytic investigation. 26. Morris EB, Zhang B, Bondy SJ. Bullying and Smoking: Examining the
Sch Psychol Q. 2010;25(2):65–83. Relationships in Ontario Adolescents. J Sch Health. 2006;76(9):465–70.
2. Leymann H. The content and development of mobbing at work. Eur J Work 27. Nansel TROM. Bullying behaviors among us youth: Prevalence and
Organ Psychol. 1996;5(2):165–84. association with psychosocial adjustment. JAMA. 2001;285(16):2094–100.
3. Due P, Holstein BE. Bullying victimization among 13 to 15 year old school 28. Moore M, Kirkham C. Self-esteem and its relationship to bullying behaviour.
children: Results from two comparative studies in 66 countries and regions. Aggress Behav. 2001;27(4):269–83.
Int J Adolesc Med Health. 2008; 20:209. 29. Marini ZA, Dane AV, Bosacki SL, Cura Y. Direct and indirect bully-victims:
4. Due P, Hansen EH, Merlo J, Andersen A, Holstein BE. Is victimization from differential psychosocial risk factors associated with adolescents involved in
bullying associated with medicine use among adolescents? A nationally bullying and victimization. Aggress Behav. 2006;32(6):551–69.
representative cross-sectional survey in Denmark. Pediatrics. 30. Brighi A, Guarini A, Melotti G, Galli S, Genta ML. Predictors of victimisation
2007;120(1):110–7. across direct bullying, indirect bullying and cyberbullying. Emotional Behav
5. Hawker DSJ, Boulton MJ. Twenty years’ research on peer victimization and Difficulties. 2012;17(3–4):375–88.
psychosocial maladjustment: A meta-analytic review of cross-sectional 31. Prinstein MJ, Boergers J, Vernberg EM. Overt and relational aggression in
studies. J Child Psychol Psychiatry. 2000;41(4):441–55. adolescents: Social–psychological adjustment of aggressors and victims.
6. Fekkes M, Pijpers FIM, Fredriks AM, Vogels T, Verloove-Vanhorick SP. Do J Clin Child Psychol. 2001;30(4):479–91.
Bullied Children Get Ill, or Do Ill Children Get Bullied? A Prospective Cohort 32. Georgiou S, Stavrinides P. Parenting at home and bullying at school.
Study on the Relationship Between Bullying and Health-Related Symptoms. Soc Psychol Educ. 2013;16(2):165–79.
Pediatrics. 2006;117(5):1568–74. 33. Holt MK, Kantor GK, Finkelhor D. Parent/child concordance about bullying
7. Kim YS, Leventhal B. Bullying and suicide. A review. Int J Adolesc Med involvement and family characteristics related to bullying and peer
Health. 2008;20(2):133–54. victimization. J Sch Violence. 2009;8(1):42–63.
8. Klomek AB, Sourander A, Kumpulainen K, Piha J, Tamminen T, Moilanen I, et 34. Smith PK, Myron-Wilson R. Parenting and School Bullying. Clin Child Psychol
al. Childhood bullying as a risk for later depression and suicidal ideation Psychiatry. 1998;3(3):405–17.
among Finnish males. J Affect Disord. 2008;109(1–2):47–55. 35. Rigby K, Cox I, Black G. Cooperativeness and bully/victim problems
9. Smith JD, Schneider BH, Smith PK, Ananiadou K. The Effectiveness of Whole- among Australian schoolchildren. J Soc Psychol. 1997;137(3):357–68.
School Antibullying Programs: A Synthesis of Evaluation Research. Sch 36. Rigby K. Bullying in schools and what to do about it(updated, revised).
Psychol Rev. 2004;33(4):547–60. ACER: Camberwell; 2007.
10. Baldry A, Farrington DP. Protective factors as moderators of risk factors in 37. McIntyre JG, Dusek JB. Perceived parental rearing practices and styles of
adolescence bullying. Soc Psychol Educ. 2005;8(3):263–84. coping. J Youth Adolesc. 1995;24(4):499–509.
11. Anderson CA, Bushman BJ. Human aggression. Annu Rev Psychol. 38. Lodge J, Feldman SS. Avoidant coping as a mediator between
2002;53(1):27–51. appearance-related victimization and self-esteem in young Australian
12. Baron RA, Richardson DR. Human aggression (2nd ed.): New York, NY, US. adolescents. Br J Dev Psychol. 2007;25(4):633–42.
New York, NY: Plenum Press; 1994. 39. Baldry AC. The Impact of Direct and Indirect Bullying on the Mental
13. Neuman JH, Baron RA. Workplace violence and workplace aggression: and Physical Health of Italian Youngsters. Aggress Behav.
Evidence concerning specific forms, potential causes, and preferred targets. 2004;30(5):343–55.
J Manag. 1998;24(3):391–419. 40. Bijttebier P, Vertommen H. Coping with peer arguments in school-age
14. Due P, Holstein BE, Lynch J, Diderichsen F, Gabhain SN, Scheidt P, et al. children with bully/victim problems. Br J Educ Psychol.
Bullying and symptoms among school-aged children: International 1998;68(3):387–94.
comparative cross sectional study in 28 countries. Eur J Pub Health. 41. Lazarus RS. Emotion and adaptation: New York, NY, US. New York, NY:
2005;15(2):128–32. Oxford University Press; 1991.
Andersen et al. BMC Psychology (2015) 3:35 Page 15 of 15

42. Antonovsky A. Unraveling the mystery of health: How people manage stress 66. Brixval CS, Rayce SLB, Rasmussen M, Holstein BE, Due P. Overweight,
and stay well: San Francisco, CA, US. San Francisco, CA: Jossey-Bass; 1987. body image and bullying. An epidemiological study of 11- to 15-years
43. Höge T, Büssing A. The Impact of Sense of Coherence and Negative olds. Eur J Pub Health. 2012;22(1):126–30.
Affectivity on the Work Stressor-Strain Relationship. J Occup Health 67. Andersen LP, Karlsen IL, Ozmec MN, Kines P, Nielsen KJ. Social identity in work
Psychol. 2004;9(3):195–205. crews at large construction sites: Implications for safety behaviour. A qualitative
44. Nielsen MB, Einarsen S. Sampling in research on interpersonal aggression. study. J of Cons Manag and Eco. 2015. doi:10.1080/01446193.2015.1087645
Aggress Behav. 2008;34(3):265–72. 68. Finkelstein DM, Kubzansky LD, Goodman E. Social Status, Stress, and
45. Hogh A, Mikkelsen EG. Is sense of coherence a mediator or moderator Adolescent Smoking. J Adolesc Health. 2006;39(5):678–85.
of relationships between violence at work and stress reactions? Scand 69. Caravita SCS, Gini G, Pozzoli T. Main and moderated effects of moral cognition
J Psychol. 2005;46(5):429–37. and status on bullying and defending. Aggress Behav. 2012;38(6):456–68.
46. Caravita SCS, Di Blasio P, Salmivalli C. Unique and interactive effects of 70. van der Watt R. Attachment, parenting styles and bullying during pubertal
empathy and social status on involvement in bullying. Soc Dev. years. J Child Adol Ment Health. 2014;26(3):251–61.
2009;18(1):140–63. 71. Bandura A, Grusec JE, Menlove FL. Observational learning as a function of
47. de Bruyn EH, Cillessen AHN, Wissink IB. Associations of peer acceptance symbolization and incentive set. Child Dev. 1966;37(3):499–506.
and perceived popularity with bullying and victimization in early 72. Strelau J, Zawadzki B. Temperament from a psychometric perspective:
adolescence. J Early Adolesc. 2010;30(4):543–66. Theory and measurement. In: Anonymous, editor. The SAGE handbook of
48. Nakamoto J, Schwartz D. Is peer victimization associated with personality theory and assessment, Vol 2: Personality measurement and
academic achievement? A meta-analytic review. Soc Dev. testing. Thousand Oaks, CA, USA: Sage Publications, Inc, Thousand Oaks, CA;
2010;19(2):221–42. 2008. p. 352–73.
49. Schwartz D, Gorman AH, Nakamoto J, Toblin RL. Victimization in the 73. Kirchner T, Forns M, Amador JA, Muñoz D. Stability and consistency
Peer Group and Children’s Academic Functioning. J Educ Psychol. of coping in adolescence: A longitudinal study. Psicothema.
2005;97(3):425–35. 2010;22(3):382–8.
50. Nansel TRCW. CRoss-national consistency in the relationship between 74. Seiffge-Krenke I, Beyers W. Coping trajectories from adolescence to young
bullying behaviors and psychosocial adjustment. Arch Pediatr Adolesc adulthood: Links to attachment state of mind. J Res Adolesc.
Med. 2004;158(8):730–6. 2005;15(4):561–82.
51. Jansen DE, Veenstra R, Ormel J, Verhulst FC, Reijneveld SA. Early risk 75. Griffith MA, Dubow EF, Ippolito MF. Developmental and cross-situational
factors for being a bully, victim, or bully/victim in late elementary differences in adolescents’ coping strategies. J Youth Adolesc.
and early secondary education. The longitudinal TRAILS study. BMC 2000;29(2):183–204.
Public Health. 2011;11:440. 76. Alsaker FD, Olweus D. Stability of global self-evaluations in early
52. Due P, Merlo J, Harel-Fisch Y, Damsgaard MT, Holstein BE, Hetland J, adolescence: A cohort longitudinal study. J Res Adolesc.
et al. Socioeconomic inequality in exposure to bullying during 1992;2(2):123–45.
adolescence: A comparative, cross-sectional, multilevel study in 35 77. Baillien E, Neyens I, De Witte H, De Cuyper N. A qualitative study on the
countries. Am J Public Health. 2009;99(5):907–14. development of workplace bullying: Towards a three way model.
53. Due P, Damsgaard MT, Lund R, Holstein BE. Is bullying equally J Community Appl Soc Psychol. 2009;19(1):1–16.
harmful for rich and poor children?: a study of bullying and 78. Ortega A, Høgh A, Pejtersen JH, Olsen O. Prevalence of workplace bullying
depression from age 15 to 27. Eur J Pub Health. 2009;19(5):464–9. and risk groups: a representative population study. Int Arch Occup Environ
54. Due P, Kroelner R, Rasmussen M, Andersen A, Damsgaard MT, Graham H, et al. Health. 2009;82(3):417–26.
Pathways and mechanisms in adolescence contribute to adult health 79. Zapf D, Einarsen S. Mobbing at Work: Escalated Conflicts in Organizations.
inequalities. Scand J Public Health. 2011;39(6):62–78. In: Anonymous, editor. Counterproductive work behavior: Investigations of
55. Glasscock D, Andersen JH, Labriola M, Rasmussen K, Hansen C. Can actors and targets. Washington, DC: American Psychological Association;
negative life events and coping style help explain socioeconomic 2005. p. 237–70.
differences in percieved stress among adolescents? A cross-sectional 80. Podsakoff PM, MacKenzie SB, Lee JY, Podsakoff NP. Common method biases
study based on the West Jutland Cohort Study. BMC Public Health. in behavioral research: A critical review of the literature and recommended
2013;13(1):1. remedies. J Appl Psychol. 2003;88(5):879–903.
56. Smith PK, Singer M, Hoel H, Cooper CL. Victimization in the school and 81. Crick NR, Bigbee MA. Relational and overt forms of peer victimization: A
the workplace: Are there any links? Br J Psychol. 2003;94(2):175–88. multiinformant approach. J Consult Clin Psychol. 1998;66(2):337–47.
82. Hansen TB, Steenberg LM, Palic S, Elklit A. A review of psychological factors
57. Schäfer M, Korn S, Smith PK, Hunter SC, Mora-Merchán J, Singer MM, et
related to bullying victimization in schools. Aggress Violent Behav.
al. Lonely in the crowd: Recollections of bullying. Br J Dev Psychol.
2012;17(4):383–7.
2004;22(3):379–94.
58. Guerra NG, Williams KR, Sadek S. Understanding Bullying and
Victimization During Childhood and Adolescence: A Mixed Methods
Study. Child Dev. 2011;82(1):295–310.
59. Goodman E, Huang B. SOcioeconomic status, depressive symptoms,
and adolescent substance use. Arch Pediatr Adolesc Med.
2002;156(5):448–53.
60. Klimidis S, Minas IH, Ata AW, Stuart GW. Construct validation in
adolescents of the Brief Current form of the Parental Bonding
Instrument. Compr Psychiatry. 1992;33(6):378–83. Submit your next manuscript to BioMed Central
61. Epstein NB, Baldwin LM, Bishop DS. The McMaster Family Assessment and take full advantage of:
Device. J Marital Fam Ther. 1983;9(2):171–80.
62. Rosenberg M, Scholler C, Schoenbach C, Rosenberg F. Global
• Convenient online submission
Self-Esteem and Specofic Self-Esteem: Different Concepts, different
outcome. Am Sociol Rev. 1995;60(1):141–56. • Thorough peer review
63. Mahammadzadeha A, Poursharifib H, Alipoura A. Validation of Sense • No space constraints or color figure charges
of Coherence (SOC) 13-item scale in Iranian sample. Procedia Soc
Behav Scie. 2010;5:1451–5. • Immediate publication on acceptance
64. Hosmer DW, Lemeshow S, Studivant R. Applied Logistic regression: • Inclusion in PubMed, CAS, Scopus and Google Scholar
3. ed. John Wiley and Sons: Hoboken, New Jersey. 2013. • Research which is freely available for redistribution
65. Giletta M, Scholte RHJ, Engels RCME, Larsen JK. Body mass index and
victimization during adolescence: The mediation role of depressive
symptoms and self-esteem. J Psychosom Res. 2010;69(6):541–7. Submit your manuscript at
www.biomedcentral.com/submit