Int Arch Occup Environ Health (2015) 88:131–142

DOI 10.1007/s00420-014-0944-7

ORIGINAL ARTICLE

Workplace bullying as an antecedent of mental health problems:

a five-year prospective and representative study
Ståle Einarsen • Morten Birkeland Nielsen

Received: 10 October 2013 / Accepted: 30 April 2014 / Published online: 20 May 2014
 Springer-Verlag Berlin Heidelberg 2014

Abstract Introduction
Purpose The present study investigates the proposed
long-term relationship between exposure to workplace Workplace bullying, in the form of long-term and sys-
bullying and subsequent mental health in the form of anx- tematic exposure to increasingly harsher forms of aggres-
iety and depression with a time lag of 5 years, exploring sion, incivility and social exclusion by other organization
potential gender differences in these relationships. members, has been theorized to mount to an ongoing
Methods The study employs a prospective design with a trauma for targets with severe long-term outcomes for
5-year time lag in a representative sample of the Norwe- those exposed (Bowling and Beehr 2006). Consistent with
gian workforce. A cohort of 1,613 employees reported on stress theory, a growing body of research has documented
their exposure to workplace bullying and their symptoms exposure to workplace bullying as a major source of dis-
of anxiety and depression at both measurement times. tress at work associated with subsequent health problems
Results The results showed exposure to workplace bul- and reduced well-being both in a shorter and longer time
lying to be a significant predictor of mental health prob- perspective (Hoobler et al. 2010). More specifically,
lems 5 years on, even after controlling for baseline mental workplace bullying has been related to reduced job satis-
health status, gender, age, job-change, job demands and job faction and engagement (Rodriguez-Munoz et al. 2009),
control, yet for men only. Baseline levels of mental health lowered commitment and turnover (Berthelsen et al. 2011;
problems in terms of symptoms of anxiety and depression Høgh et al. 2011a), as well as to increased levels of sick-
did not predict subsequent exposure to bullying at follow- ness absenteeism (Kivimäki et al. 2000; Ortega et al.
up among women, but anxiety did in the case of men. 2011), psychosomatic complaints and a range of mental
Conclusion Workplace bullying poses a serious long- health problems (Kivimäki et al. 2003; Lahelma et al.
term threat to the health and well-being of workers, at least 2012).
for men. The results of the study pinpoint the need for Despite the fact that workplace bullying is considered as
mental health treatment as well as for preventive measures a long-lasting process (Einarsen et al. 2011), most empir-
in relation to workplace bullying, and pinpoint the need for ical studies have employed cross-sectional and correla-
a gender perspective in these studies. tional research designs with only a handful studies
employing longitudinal designs (e.g., Finne et al. 2011;
Keywords Workplace bullying
Harassment
Mental Høgh et al. 2011a; Rugulies et al. 2012). Among those
health at work
Prospective design
Well-being at work
studies that have, the majority has generally employed
Social stress at work relatively short time lags of 12–24 months (see also Niel-
sen and Einarsen 2012a), which is no longer than an
S. Einarsen (&)
M. B. Nielsen average duration of a bullying case (Zapf et al. 2011).
University of Bergen, Bergen, Norway Employing a prospective research design in a representa-
e-mail: [email protected]
tive sample, the present study contributes to the literature
M. B. Nielsen by investigating the role of earlier exposure to workplace
National Institute of Occupational Health, Oslo, Norway bullying as an etiological factor for mental health problems

123
132 Int Arch Occup Environ Health (2015) 88:131–142

5 years later. As studies with shorter time lags have sug- Einarsen 2002). In line with this, clinical observations have
gested baseline mental health problems to be associated described grave outcomes of workplace bullying such as
with later exposure to bullying, thus indicating an social isolation and maladjustment, psychosomatic ill-
unhealthy or vulnerable worker mechanism or a vicious nesses, depressions, compulsions, helplessness, anger,
circle, we will also investigate the possibility of reversed anxiety and despair (Leymann 1990). Cognitive trauma
causation between the variables. theory further holds that it may not be the external event
itself that causes the trauma, but rather the more or less
Definitions and theoretical background permanent effect these events have on the perceptions and
cognitions of the target (Janoff-Bulman 1992). Hostile
The concept of workplace bullying, sometimes referred to social events or series of such events threaten our basic
as worker harassment, describes situations in the workplace cognitive schemas and fundamental beliefs about the world
where an employee persistently and over a long time per- being benevolent and meaningful, and that we, as indi-
ceives him- or herself to be mistreated and abused by other viduals, are worthy, decent and capable human beings
organization members, and where the person in question deserving affection and support by our friends and fellow
finds it difficult to defend him/herself against these actions citizens (Janoff-Bulman 1989). Under normal circum-
(Einarsen et al. 2011). While no definitive list exists of the stances, these beliefs and assumptions enable us to operate
specific behaviors involved, common acts of workplace effectively in our daily lives securing us that no harm will
bullying include angry outbursts, verbal hostility, hostile affect us, and providing us with a sense of invulnerability
acts of humor like being made a laughing stock, as well as as well as some confidence in our ability to cope with the
illegitimate and highly stressful changes in ones tasks and challenges of life (see also Einarsen and Mikkelsen 2003).
work situation such as undue work pressure, being given When facing workplace bullying, victims will instantly or
unreasonable deadlines or being exposed to a dangerous at least gradually and painfully become aware of the fra-
work situation. Quite often, bullying is conducted with acts gility of these assumptions, giving rise to a strong stress
of social exclusion or social isolation in combination with responses and possibly a long-lasting state of fear, cogni-
such direct attacks described above. Yet, it is the accu- tive confusion, anxiety and depression.
mulated number of acts over time, and the summarized As described above, the repeated hostility involved in
pattern of behaviors, and not the particular and individual bullying is often combined with acts of social exclusion
acts involved, that constitutes the menace. As isolated acts and ostracism, which in itself tend to create instant
of aggression, such incidents may be mildly offensive and symptoms of confusion and anxiety in those excluded, over
even tolerable. However, when accumulated over time, time interfering with the immune system and brain func-
these acts will be highly destabilizing and a distressing tions relating to aggression and depression (Williams
situation to those exposed (Leymann 1990). Following 1997). The underlying mechanisms are described in social
from this notion, bullying does not seem to be an either-or psychological theories where social exclusion or rejection
phenomenon, but rather a slowly escalating process with from important social relationships is claimed to threaten a
increasingly more harsh treatment of the target. Estimates range of basic social needs, including the need to belong,
based on meta-analytical evidence suggest that at least the need for a worthy self and the need to perceive the
15 % of the global working population is exposed to world as predictable and controllable. Exposure to bullying
negative behaviors in the workplace that can be experi- is assumed to profoundly question the focal persons sense
enced as bullying, whereas 11 % label themselves as vic- of being a capable and worthy individual, again reminding
tims of bullying according to the above definition (Nielsen them of their fragile existence, hence constituting a severe
et al. 2010). source of social pain and even symbolizing a kind of
Exposure to such treatment while at work has long been ‘social death’ (Mikkelsen and Einarsen 2002; Williams
claimed as a more crippling and devastating problem for 1997). As for short-term effects, evidence from a study
employees than all other work-related stress put together examining the neural correlates of social exclusion through
(Wilson 1991), with much empirical evidence as support the use of functional magnetic resonance brain imaging
(see also Høgh et al. 2011b; Nielsen and Einarsen 2012a (fMRI) showed that social pain actually activates similar
for overviews). According to the cognitive trauma theory brain bases to those of physical pain (Eisenberger et al.
of Janoff-Bulman (1992), one may deduce that victims of 2003). In the long run, perceptions of exclusion are seen to
systematic psychological hostility from other human cause extreme anxiety, depression and even psychotic
beings, be it real or perceived, react with shock, fear, reactions. Based on clinical observations, Leymann (1990),
anxiety and depression, and may even alter their percep- described how exposure to social isolation or social
tions of their surroundings and future to one of threat, ostracism gradually reduces the victim’s ability to cope
danger, insecurity and self-questioning (Mikkelsen and with the demands of daily living, reinforcing other people’s

123
Int Arch Occup Environ Health (2015) 88:131–142 133

negative attitude toward the focal person, putting gradually health, suggesting that a strain–stressor model exists parallel
more strain on the victim’s self-esteem and self-confidence. to or sequentially with the stressor–strain model. For
Generally, mainstream work stress models hold that instance, Finne et al. (2011) showed that baseline levels of
‘long-term exposure to high work stress without having mental distress predicted bullying at follow-up, with an odds
sufficient possibilities for recovery affects health in the long ratio of 2.30, suggesting a reversed causation between bul-
run via sustained activation of the autonomic nervous sys- lying and mental health problems. According to Kompier
tem and enhanced neuroendocrine responses’ (Kompier and and Taris (2011), unhealthy (e.g., depressive) employees
Taris 2011). Hence, based on such theories of work stress, evaluate their environment more negatively than do other
cognitive trauma theory and social psychological theories employees. Reported environmental ‘changes’ are therefore
on ostracism and social exclusion, we may put forward the mainly a matter of changed perceptions. Furthermore,
hypothesis that exposure to workplace bullying is associ- ‘… distressed employees may underperform and therefore
ated with subsequent elevated reports of anxiety and receive less support from colleagues or supervisors, they
depression in those targeted, even within long-time spans. may not participate in social meetings with colleagues,
In a meta-analysis of 33 cross-sectional studies on the prefer to remain alone and therefore become isolated and
relationship between bullying and mental health outcomes, ridiculed, be given the ‘silent treatment,’ and so forth—all of
it was established that the mean correlation between which may well be considered forms of bullying’ (Kompier
exposure to bullying and self-reports of lowered mental and Taris 2011). Hence, some sort of vicious circle of events
health was .34 (Nielsen and Einarsen 2012a). Employing a may exist where bullying leads to mental health problems,
prospective design with a 2-year time lag and a sample of which may act to worsen the situation for the target or at least
1971 Norwegian workers recruited from some 20 organi- worsen the perception the targets make of their work situa-
zations, Finne et al. (2011)showed that baseline exposure to tion. Employing a longer time interval between measure-
workplace bullying was associated with a subsequent ment points such as the 5 years employed in the present study
increase in mental health problems 2 years later, even when may, however, be particularly helpful in order to indicate
controlling for baseline mental distress and job demands. In where it all starts: In mental health problems or in exposure to
the aforementioned meta-analysis, a mean correlation of .20 harsh treatment by peers or superiors at work?
was revealed between exposure to workplace bullying and
subsequent reports of mental health problems in 7 studies Aim of study
employing prospective designs.
Yet, an important limitation of the existing longitudinal Few studies exist on the longitudinal relationships between
studies is that they employ relatively short or intermediate exposure to workplace bullying and mental health distress
time lags between measurement points. Most time lags are in the form of symptoms of anxiety and depression, and
only 1 or 2 years, which actually is the typical length of a those that exists mainly look at intermediate time lags of
bullying episode. Hence, little is known about the after effects 1–2 years (Nielsen and Einarsen 2012a). Yet, theoretically,
of workplace bullying over longer time-periods. A study by bullying is claimed to constitute a traumatic series of
Stoetzer et al. (2009), employing a representative sample of events with possible life-changing effects. By employing a
the Swedish work force, showed that being socially excluded prospective design with a 5-year time lag between mea-
at work was associated with an increased risk of depression surement points, the present study investigates the rela-
3 years later. Lahelma et al. (2012) found elevated risks of tionship between exposure to workplace bullying and
having a common mental disorder 5–7 years later among a subsequent mental health problems. In order to purify the
group of Finnish victims of workplace bullying, an effect that association between workplace bullying and subsequent
was strongest for male participants. mental health problems, we will control for the stability of
Yet, the mobilization–minimization hypothesis of Tay- the outcome variables, as well as other well-known
lor (1991) holds that while the effects of negative experi- stressful aspects of the target’s job role such as high
ences and events may be stronger than the effects of demands and lack of control. These control variables are
positive events in the short run, the potential effects of included as they are empirically established as important
negative experiences will quickly wane and hence be less predictors of mental health (Stansfeld and Candy 2006),
profound as time goes by, questioning whether bullying has and because workplace bullying generally takes place
such strong and long-lasting effects. Hence, more evidence in situations where employees are plagued with such
is needed on the long-term outcomes of workplace unhealthy work characteristics (Baillien et al. 2011; Hauge
bullying. et al. 2007, 2010). Hence, we investigate the long-term
In addition to the stressor–strain model described above, outcomes of bullying purified of the possible overlapping
there is also theoretical and empirical evidence for reversed effects of a general stressful work situation. We will test
associations between workplace bullying and psychological the assumption both for a compound measure of mental

123
134 Int Arch Occup Environ Health (2015) 88:131–142

distress as well as for anxiety and depression separately. In exposure to workplace bullying (Zapf and Einarsen 2011),
addition, we will investigate the possibility of reversed the exposure outcome relationships may still be different
causation between the variables, accounting for any for men and women. This possibility is surprisingly seldom
unhealthy or vulnerable worker mechanisms in explaining explored in this line of enquiry as documented in the
the occurrence of workplace bullying. The two following review by Høgh et al. (2011b) and the meta-analysis by
hypotheses will be tested: Nielsen and Einarsen (2012a). An earlier two-year pro-
spective study on the relationship between exposure to
H1 Exposure to workplace bullying is related to sub-
sexual harassment (SH) and subsequent psychological
sequent increased reports of mental health problems in the
distress did reveal such gender effects (Nielsen and Ein-
form of symptoms of anxiety and depression 5 years later,
arsen 2012b) in that SH contributed to subsequent psy-
even when controlled for base line reports of mental health
chological distress among women while psychological
problems, as well as a work situation characterized by high
distress predicted SH among men, possibly indicating
job demands and lack of control.
either a vulnerability factor or a negative perception
H2 Premorbid levels of such mental health problems are mechanism among these men.
related to later reports of being bullied at work, even when
controlled for baseline indicators of a bullying and a
stressful working environment in the form of high job Method
demands and lack of control.
Design and procedure
In line with methodological recommendations relating
to the measurement of workplace bullying (Mikkelsen and The present study is based on assessment 1 and assess-
Einarsen 2001; Nielsen et al. 2011; Salin 2003), we will ment 3 of a longitudinal survey in a nationwide sample
employ two measures of workplace bullying; systematic of the Norwegian working force. Data were collected
exposure to bullying behaviors and the perception of being with a time lag of approximately 5 years. In 2005
victimized by bullying. While the former measures expo- (baseline), a random sample of 4,500 employees was
sure to actual negative acts while at work, the latter mea- drawn from The Norwegian Central Employee Register
sures to what extent a respondent self-labels as a victim of (NCER) by Statistics Norway (SSB). The sampling cri-
bullying. By employing both measures, we may differen- teria were adults between 18 and 67 years of age that
tiate the effect of mere exposure to perceived negative acts were registered in the NCER as employed during the last
from the effect of the subjective identity of being a victim. 6 months before the survey, in an enterprise with a staff
A study based on the same baseline data as in the present of five or more, and with a mean working hour of more
study, but with a 2-year time lag between measurement than 15 h per week.
points, showed that it was the subjective feeling of being a Questionnaires were distributed through the Norwegian
victim, rather than ones actual exposure to bullying Postal Service to the respondents’ home addresses. Altogether
behaviors, that predicted subsequent increases in mental 2,539 questionnaires were returned (response rate 57 %), a
health problems (Nielsen et al. 2012). This suggests that response rate slightly higher than what is typically the case in
self-labeling as a victim may function as a mediator these kinds of workplace survey studies (Baruch and Holtom
between exposure and mental health problems. Such a 2008). Using the same procedure, the follow-up data were
finding would be in line with the cognitive trauma theory collected in 2010. All respondents from the baseline survey
and most stress theories that hold that it is not the actual were asked to participate in the follow-up. The response rate
events per se that causes the observed stress reaction, but for the follow-up survey was 64 % (N = 1,613). With the
rather the effects events have on the inner world of the exception of a somewhat skewed gender distribution, the
recipient. Yet, a study by Vie et al. (2011) showed that cohort sample can be considered as representative for the
when exposure to bullying behaviors is high, such persis- Norwegian working population with regard to demographic
tent exposure have a considerable harmful effect on the characteristics (Høstmark and Lagerstrøm 2006). The project
target’s health independently of whether the experience is was approved by the Regional Committee for Medical
labeled as bullying or not. Hence, the present study Research Ethics in Western Norway.
explores the possible different long-term relationship with
mental health problems for these two important aspects of The cohort
the bullying experience when testing the two above
hypotheses. In addition, we will explore if gender differ- Among the 1,613 respondents, women were slightly
ences may be traced in these hypothesized relationships. overrepresented (54 %). The mean age at baseline was
Although there are not necessarily gender differences in the 45.23 years. About 91 % of the respondents were

123
Int Arch Occup Environ Health (2015) 88:131–142 135

employed in a full-time (72 %) or part-time (19 %) posi- Mental health was assessed with the Hopkins Symptoms
tion, whereas the remaining respondents were on tempo- Checklist-25 (HSCL-25; Derogatis et al. 1974), which
rary sick leave, paid leave or vocational rehabilitation. A covers the most common psychiatric symptoms in the area
total of 81 % had a regular day time working arrangement. of anxiety and depression. The HSCL-25 is scored on a
Mean working hours per week was 37.3. severity scale from 1 (‘not at all’) to 4 (‘extremely’) using
the ‘last week’ as a frame of reference for answering.
Instruments Following Derogatis et al. (1974), six items assess symp-
toms of anxiety, whereas 11 items symptoms of depression.
Exposure to bullying behaviors in the workplace was Mental caseness, that is the need for treatment, is calcu-
measured with the 22-item Negative Acts Questionnaire- lated by having an average score on the HSCL-25 items
Revised (NAQ-R) inventory (Einarsen et al. 2009). NAQ-R C1.75 (Nettelbladt et al. 1993). The same cutoff value
describes behaviors that may be perceived as bullying if applies for the anxiety and depression subscales (Stammel
occurring on a regular basis. All items are formulated in et al. 2012). The comparison of scores on the HSCL-25 to
behavioral terms and hence focus on the mere exposure to physician ratings of emotional distress has yielded a con-
inappropriate behaviors while at work. The NAQ-R con- cordance rate of 86.7 % (Hesbacher et al. 1980). The
tains items referring to both direct (e.g., openly attacking HSCL-25 has satisfactory validity and reliability as a
the victim) and indirect (e.g., social isolation, slander) measure of psychological distress (Nettelbladt et al. 1993;
behaviors. It also contains items referring to personal as Strand et al. 2003). High internal consistency, as measured
well as work-related forms of bullying. For each item, the with Cronbach’s a, was found for the total scale (.92/.93),
respondents were asked how often they had been exposed as well as for the two subscales of anxiety (.79/.79) and
to the behavior at their present worksite during the last depression (.87/.88), at both baseline and follow-up.
6 months. Response categories range from 1 to 5 (‘never,’
‘now and then,’ ‘monthly,’ ‘weekly’ and ‘daily’). In Attrition analyses
accordance with recent advances on the use of cutoff cri-
teria with regard to estimating targets of bullying (No- Attrition analyses of demographic data at baseline,
telaers and Einarsen 2012), respondents with a summarized including dropouts, revealed that the cohort respondents
score of C33 on the NAQ-R were classified as targets of (M = 45.23; SD = 11.23; N = 1,613) were somewhat
workplace bullying. Cronbach’s a for the NAQ-R was high older than dropout respondents (M = 41.30; SD = 11.60;
at both baseline (.90) and follow-up (.88). N = 926) (t = -8.38; df = 2,537; p \ .001). Systematic
In line with previous studies (Nielsen et al. 2010, 2011), gender differences (v2 = 3.97; df = 1; p \ .05) were
perceived victimization from bullying was measured by revealed between cohort (54 % women) and dropouts
asking the respondents to indicate whether they considered (50 %) at baseline. No differences were established with
themselves to have been victimized by bullying at work regard to leadership responsibility, but a higher percentage
during the last 6 months according to the following defi- of the cohorts (15 %) compared with dropouts (11 %) was
nition (Einarsen and Skogstad 1996; Olweus 1991): ‘Bul- elected as a shop steward or other form of employee rep-
lying takes place when one or more persons systematically resentative. Compared with the cohort sample (11 %),
and over time feel that they have been subjected to nega- significantly more of the dropouts (15 %) reported expo-
tive treatment on the part of one or more persons, in a sure to bullying behavior at baseline (v2 = 5.77; df = 1;
situation in which the person(s) exposed to the treatment p \ .05). No differences between cohort and dropouts were
have difficulty in defending themselves against them. It is found with regard to self-labeled victimization from bul-
not bullying when two equally strong opponents are in lying. Finally, a significantly higher percentage of the
conflict with each other.’ The response categories were dropout (16 %) than among cohorts (12 %) reported
‘no,’ ‘yes, rarely,’ ‘yes, now and then,’ ‘yes, once a week’ symptoms of psychological distress above the recom-
and ‘yes, several times a week’. Respondents answering mended cutoff value of mental caseness (v2 = 9.23;
‘yes’ to this question were categorized as self-labeled df = 1; p \ .01).
victims of workplace bullying.
Eight items from the ‘Questionnaire on the Experience Statistical analysis
and Assessment of Work’ (Van Veldhoven et al. 1999)
were employed to assess job demands (Cronbach’s a .86) Statistical analyses were conducted using IBM SPSS Sta-
and control in the form of decision latitude (Cronbach’s a tistics 21.0. Logistic regression analyses were conducted in
.75). The items were scored using a 4-point scale ranging order to examine longitudinal relationships between bul-
from ‘never,’ ‘sometimes,’ ‘quite often,’ ‘through’ ‘very lying and psychological distress. The level of significance
often/nearly always.’ was set to .05. In order to determine gender differences in

123
136 Int Arch Occup Environ Health (2015) 88:131–142

Table 1 Crude (unadjusted) Baseline (T1) Follow-up (T2)

Odds ratios between baseline
predictors and measures of Psychological distress Bullying behaviors Self-labeled
distress and bullying at follow- victimization
up
OR 95 % CI OR 95 % CI OR 95 % CI

Psychological distress (no) 10.73** 7.2–15.98 3.87** 2.48–6.03 3.78** 2.04–7.00

Age .97** .95–.99 .98 .97–1.00 1.01 .98–1.03
Gender (male) 1.5* 1.06–2.13 .97 .67–1.41 .76 .44–1.30
Leadership position (no) 1.24 .79–1.93 1.22 .75–2.00 1.08 .54–2.19
Changed position (no) 1.07 .73–1.57 1.01 .66–1.54 .98 .53–1.82
Job demands 1.41* 1.05–1.90 1.53* 1.11–2.11 .96 .59–1.57
Reference categories in brackets Decision latitude .81 .61–1.08 .59** .43–.81 .73 .46–1.15
OR Odds ratio, 95 % CI 95 % Bullying behaviors (no) 4.10** 2.70–6.20 11.65** 7.63–17.79 7.85** 4.42–13.91
confidence interval
Self-labeled victimization (no) 3.37** 1.83–6.18 6.41** 3.57–11.49 11.50** 5.86–22.56
* p \ .05; ** p \ .001

the etiology of workplace bullying and psychological dis- victimization, exposure to bullying behaviors and psycho-
tress, separate analyses were conducted for men and logical distress at baseline.
women. Prospective associations between workplace bullying at
baseline and psychological distress at follow-up, adjusted
for control variables (baseline distress, age, leadership
Results position, changed position, job demands and decision lat-
itude), are presented in Table 2. In the adjusted analysis for
At baseline, 13.4 % of the cohort reported symptoms of the total sample, baseline psychological distress (OR 9.05;
psychological distress above the threshold for mental 95 % CI 5.80–14.13), being young (OR .97; 95 % CI
caseness. The corresponding number at follow-up was .95–.99) and exposure to bullying behaviors (OR 1.77;
11.4 %. At both baseline (v2 = 16.35; df = 1; p \ .001) 95 % CI 1.02–3.07) emerged as significant predictors of
and follow-up (v2 = 5.18; df = 1; p \ .05), significantly subsequent distress five years later. Based on the Cox &
more women (baseline 16.1 %/follow-up 13.2 %) than Snell R-Square and Nagelkerke pseudo R-Square values,
men (baseline 10.5 %/follow-up 9.2 %) reported symp- the predictors explained between 11 and 21 % of the var-
toms of distress above threshold for caseness. iance in psychological distress. The regression model was
A total of 12.5 % of respondents at baseline, and 9.2 % supported by a significant chi-square test (v2 = 14.17;
at follow-up, reported exposure to workplace bullying df = 8; p \ .001) and by a nonsignificant Hosmer and
behaviors above the cutoff value suggested by Notelaers Lemeshow test (v2 = 5.54; df = 8; p [ .05). Yet, gender
and Einarsen (2012). In total, 4.6 % labeled themselves as differences in the etiology may exist, as a significant
victims of bullying at baseline, while 4.1 % self-labeled at association between bullying behaviors at baseline and
follow-up. No gender differences were found with regard subsequent psychological distress was only found among
to the two indicators of bullying. men (OR 4.22; 95 % CI 1.91–9.32), and not among women
Bivariate, unadjusted associations between predictor (OR .90; 95 % CI .40–2.02).
variables at baseline and psychological distress, exposure Table 3 displays longitudinal effects of baseline psy-
to bullying behaviors, and self-labeled victimization from chological distress on exposure to bullying behaviors at
bullying at follow-up for the total sample are presented in follow-up adjusted for control variables at baseline. When
Table 1. The findings show that baseline distress, being adjusting for baseline bullying behaviors (OR 9.36; 95 %
young, being female, having high levels of job demands, CI 5.54–15.80) and the other control variables, psycho-
reporting high exposure to bullying behaviors, and claim- logical distress was not related to exposure to bullying
ing victimization from bullying, all predicted subsequent behaviors at follow-up (OR 1.55; 95 % CI .90–2.69) in the
distress, with baseline distress and the two measures of total sample. Similarly, no association was established
bullying as the strongest predictors. Exposure to bullying between self-labeled victimization at baseline and sub-
behaviors at follow-up was predicted by baseline exposure sequent exposure to workplace bullying behaviors (OR
to bullying behavior, self-labeled victimization, psycho- 1.61; 95 % CI .78–3.30) in the total sample. The predictors
logical distress, job demands and decision latitude. Sub- explained between 10 and 21 % of the variance in exposure
sequent self-labeled victimization was predicted by earlier to bullying behaviors. The tests for model fit supported the

123
Int Arch Occup Environ Health (2015) 88:131–142 137

Table 2 Longitudinal Baseline variables Men Women Total

relationships between baseline
workplace bullying and OR 95 % CI OR 95 % CI OR 95 % CI
psychological distress follow-up for OR for OR for OR

Psychological distress (no) 6.47** 3.05–13.73 12.47** 6.88–22.61 9.05** 5.80–14.13

Age .97 .94–1.01 .96* .94–.99 .97* .95–.99
Leadership position (no) .91 .45–1.84 1.68 .74–3.81 1.15 .69–1.91
Changed position (no) .97 .48–1.96 .52* .28–.98 .71 .45–1.13
Reference categories in
brackets. Age, job demands and Job demands 1.52 .88–2.60 .89 .56–1.41 1.13 .80–1.59
decision latitude are continuous Decision latitude 1.60 .92–2.79 1.07 .69–1.67 1.17 .84–1.64
variables Bullying behaviors (no) 4.22** 1.91–9.32 .90 .40–2.02 1.77* 1.02–3.07
OR Odds ratio, 95 % CI 95 % Self-labeled victimization (no) 1.57 .51–4.89 1.04 .33–3.23 1.34 .60–2.97
confidence interval
Constant .02* .21 .12*
* p \ .05; ** p \ .001

Table 3 Longitudinal Baseline variables Men Women Total

relationships between baseline
psychological distress and OR 95 % CI OR 95 % CI OR 95 % CI
exposure to bullying behaviors for OR for OR for OR
at follow-up
Bullying behaviors (no) 17.88** 7.98–40.5 5.61** 2.71–11.61 9.36** 5.54–15.80
Age 1.00 .97–1.03 .99 .96–1.02 1.00 .97–1.02
Leadership position (no) 1.43 .62–3.30 1.09 .49–2.43 1.23 .70–2.17
Changed position (no) .74 .33–1.66 .58 .28–1.19 .62 .37–1.06
Reference categories in
brackets. Age, job demands and Job demands .76 .40–1.43 1.19 .72–1.97 .98 .66–1.44
decision latitude are continuous Decision latitude .99 .54–1.82 .89 .55–1.43 .89 .62–1.30
variables Psychological distress (no) 2.92* 1.23–6.96 1.11 .53–2.32 1.55 .90–2.69
OR Odds ratio, 95 % CI 95 % Self-labeled victimization (no) 1.17 .38–3.59 2.05 .79–5.34 1.61 .78–3.30
confidence interval
Constant .11* .09 .07*
* p \ .05;** p \ .001

model with a significant chi-square test (v2 = 124.27; between baseline psychological distress and victimization
df = 8; p \ .001) and an insignificant Hosmer and Leme- at follow-up. However, exposure to bullying behaviors at
show test (v2 = 9.97; df = 8; p [ .05). Again, gender baseline predicted subsequent victimization among men
differences were found as exposure to bullying at follow-up (OR 9.93; 95 % CI 3.13–31.51), but not women (OR 1.61;
was significantly predicted by psychological distress 95 % CI .78–3.30).
among men (OR 2.92; 95 % CI 1.23–6.96), but not women
(OR 1.11; 95 % CI .53–2.32). Relationships between bullying and anxiety
The adjusted relationship between baseline psychologi- and depression subscales
cal distress and self-labeled victimization at follow-up is
presented in Table 4. For the total sample, the findings In order to further add to the understanding of how bullying
show that self-labeled victimization at follow-up is pre- is related to psychological distress, supplementary analyses
dicted by baseline victimization (OR 4.58; 95 % CI were conducted for the HSCL-25 subscales of anxiety and
1.91–10.97) and baseline exposure to bullying behaviors depression. In line with the main analyses, all supple-
(OR 5.32; 95 % CI 2.40–11.79), but not baseline psycho- mentary analyses were adjusted for the stability of the
logical distress (OR 1.27; 95 % CI .57–2.85). In this outcome variable baseline from baseline to follow-up, age,
model, the predictors explained between 4 % and 15 % of leadership responsibility, changed position, job demands
the variance in subsequent victimization from bullying. A and decision latitude.
significant chi square (v2 = 51.76; df = 8; p \ .001) and a In the analyses of relationship between baseline bullying
nonsignificant Hosmer and Lemeshow test (v2 = 7.23; and anxiety at follow-up in the total sample, significant
df = 8; p [ .05) yielded support to the model. No gender associations with the outcome variable emerged for base-
differences were found with regard to the association line anxiety (OR 9.32; 95 % CI 5.70–15.23) and exposure

123
138 Int Arch Occup Environ Health (2015) 88:131–142

Table 4 Longitudinal Baseline variables Men Women Total

relationships between baseline
psychological distress and self- OR 95 % CI for OR 95 % CI for OR 95 % CI for
labeled victimization from OR OR OR
bullying at follow-up
Self-labeled victimization 6.77** 1.89–24.28 9.36* 5.54–15.80 4.58** 1.91–10.97
(no)
Age 1.01 .97–1.06 1.00 .97–1.02 1.01 .98–1.04
Leadership position (no) .85 .29–2.43 1.23 .70.–2.17 .92 .42–2.01
Changed position (no) .46 .14–1.56 .62 .37–1.06 .56 .25–1.24
Reference categories in
brackets. Age, job demands and Job demands .32* .12–.84 .98 .66–1.44 .56 .31–1.01
decision latitude are continuous Decision latitude .78 .33–1.85 .89 .62–1.30 .97 .56–1.67
variables Psychological distress (no) .55 .14–2.25 1.55 .90–2.69 1.27 .57–2.85
OR Odds ratio, 95 % CI 95 % Bullying behaviors (no) 9.93** 3.13–31.51 1.61 .78–3.30 5.32 2.40–11.79
confidence interval
Constant .58 .07* .10
** p \ .001

to bullying behaviors (OR 2.15; 95 % CI 1.17–3.94). The CI .54–3.23 and depression (OR 1.31; 95 % CI .57–3.00)
other included control variables, as well as self-labeled was not related to subsequent victimization from bullying.
victimization from bullying (OR 2.03; 95 % CI .89–4.59), No significant gender differences were found for this
was not related to subsequent anxiety. In the separate association.
analyses of male respondents, anxiety at follow-up was
significantly predicted by anxiety (OR 8.23; 95 % CI
3.90–17.40), exposure to bullying behaviors (OR 3.87; Discussion
95 % CI 1.65–9.06) and victimization (OR 3.11; 95 % CI
1.00–9.65) at baseline. Baseline anxiety (OR 12.52; 95 % Exposure to workplace bullying has been claimed to be a
CI 6.23–25.14) emerged as the only significant predictor of serious social stressor with detrimental long-term health
anxiety at follow-up among women. consequences for those exposed (see Høgh et al. 2011b for
As for the relationship between bullying and subsequent a review). Still, as the vast majority of existing longitudinal
symptoms of depression in the total sample, significant studies on outcomes of workplace bullying have been
associations was found for age (OR .97; 95 % CI .95–.99) based on relatively short time lags (\ 24 months), the
and baseline depression (OR 9.60; 95 % CI 6.38–14.44), empirical basis for real long-term consequences of expo-
whereas the two indicators of bullying were not related to sure to bullying is relatively scarce. In order to add to the
depression. When separating on gender, depression at fol- understanding of how bullying is related to individual
low-up was significantly predicted by depression at base- health and well-being over longer time-periods, the aim of
line (OR 12.34; 95 % CI 6.21–24.51) and baseline the current study was to determine associations between
victimization from bullying (OR 4.12; 95 % CI bullying and psychological distress in the form of anxiety
1.37–12.39) among men. Neither exposure to bullying nor and depression in a representative sample of the Norwegian
victimization from bullying was related to depression workforce with a five year time lag between measurement
among female respondents. points.
With regard to the relationships between anxiety and The results showed that exposure to bullying behaviors
depression at baseline and the indicators of bullying at while at work predicted elevated levels of distress even five
follow-up, findings showed that exposure to bullying years later. This main finding is in line with the rather
behaviors at follow-up was significantly predicted by limited body of existing evidence showing that exposure to
exposure to bullying behaviors (OR 9.08; 95 % CI bullying is associated with psychological distress in a long
5.42–15.21) and anxiety (OR 1.89; 95 % CI 1.03–3.46), timeframe (Høgh et al. 2005; Lahelma et al. 2012).
but not depression (OR 1.30; 95 % CI .73–2.34), at base- Extending previous research, the findings of this study
line. No gender differences were established for the rela- further showed that the association between exposure to
tionship between anxiety and depression at baseline and bullying behaviors and subsequent psychological distress
exposure to bullying behaviors at follow-up. persisted over a five year period, even after controlling for
Victimization from bullying at follow-up was predicted the subjective feeling of being victimized by bullying and
by victimization (OR 4.36; 95 % CI 1.84–10.46), and for having a generally stressful work situation plagued with
exposure to bullying behaviors (OR 5.13; 95 % CI high demands and low control. As a second contribution,
2.34–11.23) at baseline. Baseline anxiety (OR 1.33; 95 % our findings indicated gender differences in that the

123
Int Arch Occup Environ Health (2015) 88:131–142 139

relationship between bullying and later symptoms of dis- problems in a shorter time-frame, whereas in a longer
tress was valid among male, but not female, respondents. timeframe, the effect of distress on bullying diminishes and
As for the specific indicators of distress, the findings disappears for females, but not necessarily for men. While
showed that perceived exposure to actual demeaning anxious men seem to have the tendency to report elevated
behaviors predict both subsequent anxiety and depression, levels of exposure to bullying behaviors even as long as
whereas the perceived identity as a victim of bullying only five years down the road, no such relationship is found
predicted subsequent levels of depressive symptoms, and among women. An implication of this finding is that it
among male respondents only. The results are thereby questions the notions about a vicious circle between bul-
leaning toward Vie et al. (2011) who found that self- lying and distress, which has been set forward in previous
labeling as a victim did not add to the explained variance in research (Kivimäki et al. 2003; Nielsen and Einarsen
mental health under conditions of high exposure to bully- 2012a). Specifically, our findings indicate that a reversed
ing behaviors. effect of distress on bullying may only be valid for a
Taken together, these findings partly support the first limited time period for women, whereas psychological
study hypothesis, which claimed that exposure to bullying distress in the form of anxiety increases the risk of being
is associated with subsequent distress five years later. This bullied in the long run among men as discussed above.
evidence for a stressor-strain model, where exposure to Yet, given the long time between measurement points,
bullying behavior influences subsequent symptoms of we do not know exactly when bullying affected the health
anxiety and depression among these targeted men, is in line of those exposed or when existing health problems influ-
with general stress theories, as well as the Cognitive ence the likelihood of being bullied. What we do know
trauma theory of Janoff-Bulman (1992) and social psy- from the results of the present study is that any longitudinal
chology theories on ostracism and social exclusion. For relationship between bullying and mental health is trace-
instance, following the basic tenets of the cognitive acti- able five years later, at least among male targets. Hence,
vation theory of stress (CATS) by Ursin and Eriksen the individual consequences of bullying do not necessarily
(2004), repeated or chronic cognitive activation created by disappear with time. We furthermore know that it is not so
stressors, such as bullying, may lead to prolonged physi- that bullying is merely a part of a general stressful work
ological activation, which again may lead to impairment in situation, nor can the relationship be explained by the
health, especially when coping efforts fails. subjective identity of someone being a victim of bullying.
As for our second hypothesis, which proposed that However, as a reversed model only received partial sup-
psychological distress at baseline is associated with an port, and since we controlled for baseline levels of psy-
elevated risk of subsequent exposure to bullying, the chological distress, we can rule out some of the impact of a
findings from the overall sample provided no overall evi- vulnerable worker effect or a gloomy perception mecha-
dence for this kind of strain-stressor model in a five-year nism in the long run, although such an effect may exist
perspective. Yet, when analyzing men and women sepa- among male targets in relation to anxiety.
rately, baseline anxiety predicted subsequent exposure to To add to our knowledge on short- and long-term indi-
bullying among males. Hypothesis 2 was therefore rejected vidual outcomes of exposure to bullying, future studies
for women but not for men, thus suggesting that the idea of should employ longitudinal designs with multiple mea-
a gloomy perception mechanism or a vulnerable worker surement points and with samples large enough to look
mechanism within such long-time frames (de Lange et al. more closely into potential gender differences. Such studies
2005; Kompier and Taris 2011) may exist among men. may also add to our knowledge on the mechanisms
This may be in line with earlier studies showing that dis- involved in this observed relationship. In this large scale
tressed males elicit more rejection from others than do samples are needed as bullying is a low-base rate phe-
females (see Segrin 1998). nomenon in most countries. With regard to mechanisms,
Earlier studies on reversed relationships between we suggest that the Cognitive Trauma Theory of Janoff-
symptoms of distress and bullying have not investigated Bulman (1992) could account for the observed relationship
gender differences and have mainly employed time lags of by predicting shattered cognitions and basic assumptions
1–2 years (Finne et al. 2011; Kivimäki et al. 2003; Rugu- among targets of bullying (see Mikkelsen and Einarsen
lies et al. 2012), something which coincide with the typical 2002), whereas the Affect Events Theory (AET; Weiss and
length of a bullying case. Hence, our findings advance Cropanzano 1996) can explain the relationship by showing
previous research by adding to the understanding of how that bullying leads to psychological distress via an increase
timeframes and gender influences the etiology of work- in one’s negative emotional state. This latter perspective, as
place bullying. That is, taking findings from all studies well as general stress models, such as the CATS (Ursin and
together, the total evidence suggests that there may be a Eriksen 2004), is supported by a cross-sectional study by
reversed causation between bullying and mental health Vie et al. (2012) where it was showed that negative

123
140 Int Arch Occup Environ Health (2015) 88:131–142

emotions and in particular, the feeling of being stressed have a stronger tendency to go on sick leave as well as
mediated the relationship between bullying and mental being on disability benefits than do men, a potential
health outcomes, ‘healthy worker effect’ may be somewhat stronger for
women than for men, potentially explaining why our
findings mainly hold for men.
Limitations

The present study poses some notable strengths. It employs Conclusion and practical implications
well-established and validated measures, testing its
hypotheses in a representative nationwide sample with a In conclusion, the present study shows that exposure to
prospective design and with response rates above the mean workplace bullying is related to psychological distress five
of organizational surveys in general (Baruch and Holtom years later even when controlling for base line levels of
2008). Hence, our findings should be seen as more robust distress and a general stressful work situation. Yet, gender
and valid than are those from cross-sectional studies who differences seem to exist, as an association between
have dominated this field for many years (see Nielsen and exposure to bullying and symptoms of distress only holds
Einarsen 2012a for an overview). The findings may also be for male targets. As for the reversed association between
generalized to a wide working population. The long time distress and bullying, the study findings indicate that
lag between measurement points, as well as the inclusion baseline anxiety is related to later exposure to bullying, yet
of control variables, such as job demands and control, puts among male respondents only. Taken together, the findings
the hypotheses on a strong test, in line with the theoretical suggest a reciprocal relationship between bullying and
and conceptual notion that bullying may be a long-lasting anxiety among male respondents, although the strength of
and life-changing situation for many of those exposed and the associations seems to indicate that that the long-term
with long-term detrimental outcomes. causal path between bullying and distress goes from the
Yet, some limitations should also be mentioned. First of former to the latter, in line with theoretical frameworks on
all, all variables are based on self-report. Hence, we do not the outcomes of trauma and social exclusion. Furthermore,
know the objective level of exposure to bullying, nor do we in a long-term perspective, it seems to be the perception of
have objective markers of mental health problems such as the mere bullying behaviors that matter the most for sub-
medically certified diagnoses. Furthermore, the observed sequent mental health and somewhat less the subjective
relationship between bullying and subsequent mental identity of being a victim of such bullying. In this, the
health problems may be inflated due to problems with causal path between the two aspects of bullying goes from
common method variance (Podsakoff et al. 2003). Yet, the exposure to self-labeling, suggesting that it all starts with
long time lag between the measurement points would tend exposure to bullying behaviors.
to reduce this problem. Furthermore, these effects have The finding of a stressor-strain, and less so a strain-
been questioned based on empirical evidence raising doubt stressor relationship between exposure to bullying and
about whether the method itself produces systematic vari- mental health, has important implications for preventive
ance in observations that inflates correlations to any sig- measures against workplace bullying and for treatment of
nificant degree (Spector 2006). Although the sample victims. First of all, in line with studies questioning the role
employed is a large one, the number of targets is still rel- of individual characteristics as an antecedent of workplace
atively few, particularly when investigating men and bullying (Glasø et al. 2009; Lind et al. 2009), the lesser
females separately. Confirmation of the null hypothesis is evidence of a strain-stressor association shows that pre-
difficult in such small samples. Therefore, the gender dif- ventive measures against bullying should be directed
ferences found in the present study should be seen as against situational factors rather than individual ones. That
mainly tentative and indicative and followed up in studies is, as long as mental health problems to a lesser extent
with an even larger sample. predict subsequent bullying than vice verse, there must be
In line with this, as the attrition analyses showed that other factors that cause bullying. Similarly, with regard to
targets of bullying tended to participate less in the follow- the treatment of victims, it is important for practitioners to
up than did other respondents, the observed relationship make victims understand that their exposure to bullying has
between bullying and mental health symptoms may even little to do with them as persons. Hence, following the
be somewhat underestimated, a tendency that may be fur- aforementioned cognitive theory of trauma, victims should
ther strengthen by the fact that we now look at rather be able to rebuild their assumptions about themselves, the
healthy employees who are still active in working life five world and others in a positive manner, something which
years later, of which some have faced other stressors and may help them deal with the bullying. Yet, both managers
health problems than bullying along the way. As females and therapist should take note of the fact of earlier studies

123
Int Arch Occup Environ Health (2015) 88:131–142 141

that within an ongoing bullying case, the reactions of tar- Eisenberger NI, Lieberman MD, Williams KD (2003) Does rejection
gets may fuel back and lead to more bullying or at least to a hurt? An fMRI study of social exclusion. Science 302:290–292
Finne LB, Knardahl S, Lau B (2011) Workplace bullying and mental
stronger perception of being bullied. Hence, further vic- distress—a prospective study of Norwegian employees. Scand J
timization may be prevented by addressing this among both Work Environ Health 37(4):276–286
targets and potential perpetrators. Glasø L, Nielsen MB, Einarsen S (2009) Interpersonal problems
among targets and perpetrators of workplace bullying. J Appl
Acknowledgments The present study was based on data from a Soc Psychol 39(6):1316–1333
collaborative project between the University of Bergen and Statistics Hauge LJ, Skogstad A, Einarsen S (2007) Relationships between
Norway (Statistisk Sentralbyrå), where the latter institution drew the stressful work environments and bullying: results of a large
sample and collected the data. The project was made possible by representative study. Work Stress 21(3):220–242
grants from two Norwegian Employer Associations (Næringslivets Hauge LJ, Skogstad A, Einarsen S (2010) The relative impact of
Hovedorganisasjon and Kommunenes Sentralforbud) and the Nor- workplace bullying as a social stressor at work. Scand J Psychol
wegian Government (Rikstrygdeverket/NAV Farve). Thanks to Bengt 51(5):426–433
Oscar Lagerstrøm, Maria Høstmark and Aina Holmøy in Statistics Hesbacher PT, Rickels K, Morris R, Newman H, Rosenfield H (1980)
Norway, and Stig Berge Matthiesen and Anders Skogstad at the Psychiatric illness in family practice. J Clin Psychiatry 41:6–10
Department of Psychosocial Science at the University of Bergen for Høgh A, Henriksson ME, Burr H (2005) A 5-year follow-up study of
their contribution. We thank the anonymous reviewers for their aggression at work and psychological health. Int J Behav Med
valuable comments on earlier drafts of this article. 12(4):256–265
Høgh A, Hoel H, Carneiro IG (2011a) Bullying and employee
Conflict of interest Ståle Einarsen runs a part-time business turnover among healthcare workers: a three-wave prospective
delivering anti-bullying training for managers and health and safety study. J Nurs Manag 19:742–751
personnel. Morten B. Nielsen declares no conflict of interest. Høgh A, Mikkelsen EG, Hansen ÅM (2011b) Individual conse-
quences of workplace bullying/mobbing. In: Einarsen S, Hoel H,
Zapf D, Cooper CL (eds) Bullying and harassment in the
workplace: developments in theory, research, and practice, 2nd
References edn. CRC Press, Boca Raton, pp 107–128
Hoobler JM, Rospenda KM, Lemmon G, Rosa JA (2010) A within-
Baillien E, Rodriguez-Munoz A, de Witte H, Notelaers G, Moreno- subject longitudinal study of the effects of positive job experi-
Jimenez B (2011) The demand-control model and target’s ences and generalized workplace harassment on well-being.
reports of bullying at work: a test within Spanish and Belgian J Occup Health Psychol 15(4):434–451. doi:10.1037/A0021000
blue-collar workers. Eur J Work Org Psychol 20(2):157–177 Høstmark M, Lagerstrøm BO (2006) Undersøkelse om arbeidsmiljø:
Baruch Y, Holtom BC (2008) Survey response rate levels and trends Destruktiv atferd i arbeidslivet. Dokumentasjonsrapport (A study
in organizational research. Hum Relat 61(8):1139–1160 of work environments: destructive behaviours in working life.
Berthelsen M, Skogstad A, Lau B, Einarsen S (2011) Do they stay or Documentation report). Statistisk Sentralbyrå/Statistic Norway,
do they go? A longitudinal study of intentions to leave and Oslo, p 42
exclusion from working life among targets of workplace Janoff-Bulman R (1989) Assumptive worlds and the stress of
bullying. Int J Manpower 32(2):178–193. doi:10.1108/ traumatic events: applications of the schema construct. Soc
01437721111130198 Cogn 4:113–136
Bowling NA, Beehr TA (2006) Workplace harassment from the Janoff-Bulman R (1992) Shattered assumptions. Towards a new
victim’s perspective: a theoretical model and meta analysis. psychology of trauma. The Free Press, New York
J Appl Psychol 91(5):998–1012 Kivimäki M, Elovainio M, Vathera J (2000) Workplace bullying and
de Lange AH, Taris TW, Kompier MAJ, Houtman ILD, Bongers PM sickness absence in hospital staff. Occup Environ Med 57:656–660
(2005) Different mechanisms to explain the reversed effects of Kivimäki M, Virtanen M, Vartia M, Elovainio M, Vathera J,
mental health on work characteristics. Scand J Work Environ Keltikangas-Järvinen L (2003) Workplace bullying and the risk
Health 31(1):3–14 of cardiovascular disease and depression. Occup Environ Med
Derogatis LR, Lipman RS, Rickels K, Uhlenhuth EH, Covi L (1974) 60:779–783
The Hopkins Symptom Checklist (HSCL): a self report symptom Kompier MAJ, Taris TW (2011) Understanding the causal relations
inventory. Behav Sci 19(1):1–15 between psychosocial factors at work and health—a circular
Einarsen S, Mikkelsen EG (2003) Individual effects of exposure to process. Scand J Work Environ Health 37(4):259–261. doi:10.
bullying at work. In: Einarsen S, Hoel H, Zapf D, Cooper CL 5271/Sjweh.3172
(eds) Bullying and emotional abuse in the workplace: interna- Lahelma E, Lallukka T, Laaksonen M, Saastamoinen P, Rahkonen O
tional perspectives in research and practice. Taylor & Francis, (2012) Workplace bullying and common mental disorders: a
London, pp 127–144 follow-up study. J Epidemiol Community Health 66(6):e3.
Einarsen S, Skogstad A (1996) Bullying at work: epidemiological doi:10.1136/jech.2010.115212
findings in public and private organizations. Eur J Work Org Leymann H (1990) Mobbing and psychological terror at workplaces.
Psychol 5:185–201 Violence Vict 5(2):119–126
Einarsen S, Hoel H, Notelaers G (2009) Measuring bullying and Lind K, Glasø L, Pallesen S, Einarsen S (2009) Personality profiles
harassment at work: validity, factor structure, and psychometric among targets and nontargets of workplace bullying. Eur
properties of the Negative Acts Questionnaire—revised. Work Psychol 14(3):231–237. doi:10.1027/1016-9040.14.3.231
Stress 23(1):24–44 Mikkelsen EG, Einarsen S (2001) Bullying in Danish worklife:
Einarsen S, Hoel H, Zapf D, Cooper CL (2011) The concept of prevalence and health correlates. Eur J Work Org Psychol
bullying and harassment at work: the European tradition. In: 10:393–414
Einarsen S, Hoel H, Zapf D, Cooper CL (eds) Bullying and Mikkelsen EG, Einarsen S (2002) Basic assumptions and symptoms
harassment in the workplace: developments in theory, research, of post-traumatic stress among victims of bullying at work. Eur J
and practice, 2nd edn. CRC Press, Boca Raton, pp 3–40 Work Org Psychol 11:87–111

123
142 Int Arch Occup Environ Health (2015) 88:131–142

Nettelbladt P, Hansson L, Stefansson C-G, Borgquist L, Nordström G Spector PE (2006) Method variance in organizational research—truth
(1993) Test characteristics of the Hopkins Symptom Check List- or urban legend? Organ Res Methods 9(2):221–232. doi:10.
25 (HSCL-25) in Sweden, using the Present State Examination 1177/1094428105284955
(PSE-9) as a caseness criterion. Soc Psychiatry Psychiatr Stammel N, Neuner F, Bottche M, Knaevelsrud C (2012) Construc-
Epidemiol 28:130–133 tion of a questionnaire for readiness to reconcile in victims of
Nielsen MB, Einarsen S (2012a) Outcomes of workplace bullying: a human rights violations. Eur J Psychotraumatol 3. doi:10.3402/
meta-analytic review. Work Stress 26(4):309–332 ejpt.v3i0.15785
Nielsen MB, Einarsen S (2012b) Prospective relationships between Stansfeld S, Candy B (2006) Psychosocial work environment and
workplace sexual harassment and psychological distress. Occup mental health—a meta-analytic review. Scand J Work Environ
Med (Lond) 62(3):226–228. doi:10.1093/occmed/kqs010 Health 32(6):443–462
Nielsen MB, Matthiesen SB, Einarsen S (2010) The impact of Stoetzer U et al (2009) Problematic interpersonal relationships at
methodological moderators on prevalence rates of workplace work and depression: a Swedish prospective cohort study.
bullying. A meta-analysis. J Occup Org Psychol 83(4):955–979. J Occup Health 51(2):144–151
doi:10.1348/096317909X481256 Strand BH, Dalgard OS, Tambs K, Rognerud M (2003) Measuring the
Nielsen MB, Notelaers G, Einarsen S (2011) Measuring exposure to mental health status of the Norwegian population: a comparison
workplace bullying. In: Einarsen S, Hoel H, Zapf D, Cooper CL of the instruments SCL-25, SCL-10. SCL-5 and MHI-5 (SF-36).
(eds) Bullying and emotional abuse in the workplace develop- Nordic J Psychiatry 57:113–118
ments in theory, research and practice. CRC Press, Boca Raton Taylor SE (1991) Asymmetrical effects of positive and negative
Nielsen MB, Hetland J, Matthiesen SB, Einarsen S (2012) Longitu- events—the mobilization minimization hypothesis. Psychol Bull
dinal relationships between workplace bullying and psycholog- 110(1):67–85. doi:10.1037//0033-2909.110.1.67
ical distress. Scand J Work Environ Health 38(1):38–46. doi:10. Ursin H, Eriksen HR (2004) The cognitive activation theory of stress.
5271/sjweh.3178 Psychoneuroendocrinology 29(5):567–592. doi:10.1016/S0306-
Notelaers G, Einarsen S (2012) The world turns at 33 and 45: defining 4530(03)00091-X
simple cutoff scores for the Negative Acts Questionnaire- Van Veldhoven M, Broesen JPJ, Fortuin RJ (1999) Werkstress in
Revised in a representative sample. Eur J Work Org Psychol beeld. Psychososiale arbeidsbelasting en werkstress in Nederland
(IFirst article) (Focus on stress. Psychososial workload and work stress in the
Olweus D (1991) Bullying/victim problem among school children. In: Netherlands). Stichting Kwaliteitsbevordering Bedriftsgezond-
Rubin I, Pepler D (eds) The development and treatment of heidszorg, Amsterdam
childhood aggression. Erlbaum, Hillsdale Vie TL, Glasø L, Einarsen S (2011) Health outcomes and self-
Ortega A, Christensen KB, Høgh A, Rugulies R, Borg V (2011) One- labeling as a victim of workplace bullying. J Psychosom Res
year prospective study on the effect of workplace bullying on 70(1):37–43. doi:10.1016/j.jpsychores.2010.06.007
long-term sickness absence. J Nurs Manag 19(6):752–759. Vie TL, Glaso L, Einarsen S (2012) How does it feel? Workplace
doi:10.1111/j.1365-2834.2010.01179.x bullying, emotions and musculoskeletal complaints. Scand J
Podsakoff PM, MacKenzie SB, Lee JY, Podsakoff NP (2003) Psychol 53(2):165–173. doi:10.1111/j.1467-9450.2011.00932.x
Common method biases in behavioral research: a critical review Weiss HM, Cropanzano R (1996) Affective events theory: a
of the literature and recommended remedies. J Appl Psychol theoretical discussion of the structure, causes and consequences
88(5):879–903. doi:10.1037/0021-9101.88.5.879 of affective experiences at work. In: Staw BM, Cummings LL
Rodriguez-Munoz A, Baillien E, De Witte H, Moreno-Jimenez B, (eds) Research in organizational behavior: an annual series of
Pastor JC (2009) Cross-lagged relationships between workplace analytical essays and critical reviews, vol 18. Elsevier Science/
bullying, job satisfaction and engagement: two longitudinal JAI Press, US, pp 1–74
studies. Work Stress 23(3):225–243. doi:10.1080/ Williams KD (1997) Social ostracism. In: Kowalski RM (ed)
02678370903227357 Aversive interpersonal behaviors. Plenum, New York,
Rugulies R et al (2012) Bullying at work and onset of a major pp 133–170
depressive episode among Danish female eldercare workers. Wilson CB (1991) US business suffer from workplace trauma. Pers
Scand J Work Environ Health 38(3):218–227. doi:10.5271/ J(July):47–50
Sjweh.3278 Zapf D, Einarsen S (2011) Individual antecedents of bullying: victims
Salin D (2003) Workplace bullying among business professionals. and perpetrators. In: Einarsen S, Hoel H, Zapf D, Cooper CL
Prevalence, organisational antecedents and gender differences. (eds) Bullying and harassment in the workplace. CRC Press,
Swedish School of Economics and Business Administration Boca Raton, pp 177–200
Segrin C (1998) Interpersonal communication problems associated Zapf D, Einarsen S, Hoel H, Vartia M (2011) Empirical findings on
with depression and loneliness. In: Andersen PA, Guerrero LK prevalence and risk groups of bullying in the workplace. In:
(eds) Handbook of communication and emotion: research, Einarsen S, Hoel H, Zapf D, Cooper C (eds) Bullying and
theory, applications, and contexts. Academic Press, San Diego, emotional abuse in the workplace developments in theory,
CA, pp 215–242 research and practice. CRC Press, Boca Raton, pp 75–105

123