You are on page 1of 16

International

Journal of Organisational, work group


Manpower
20,1/2
related and personal causes of
mobbing/bullying at work
70 Dieter Zapf
Johann Wolfgang Goethe-UniversitaÈt, Frankfurt, Germany
Keywords Bullying, Organizational behaviour, Organizational conflict
Abstract This article investigates the causes of mobbing (bullying) at work. Mobbing is defined
as a severe form of social stressors at work. Unlike ``normal'' social stressors, mobbing is a long
lasting, escalated conflict with frequent harassing actions systematically aimed at a target person.
It is argued that the organisation, the social system, a certain perpetrator and the victim have to
be considered as potential causes of mobbing. Results of two samples of mobbing victims and a
control group support this view. It is concluded that one-sided explanations on the causes of
mobbing are likely to be inappropriate and that many cases are characterised by multi-causality ±
a common finding in conflict research.

Introduction
Based on research in Sweden, Heinz Leymann (1993a; 1993b; 1996) introduced
the concept of mobbing (bullying) as a severe form of harassment in
organisations. Mobbing is defined as psychological aggression that often
involves a group of ``mobbers'' rather than a single person. Theoretically,
mobbing is an extreme type of social stressor at work. Unlike ``normal'' social
stressors, however, mobbing is a long-lasting, escalated conflict with frequent
harassing actions systematically aimed at a target person (Einarsen and
Skogstad, 1996; Leymann, 1996; Zapf, 1999; Zapf et al., 1996). Bullying, on the
other hand, connotes physical aggression by a single person, mostly by a
supervisor. Empirical studies have shown that this is clearly not the case (Zapf,
1999). In Germany, the term ``mobbing'' is preferred, but in this article the terms
mobbing and bullying are used interchangeably.
The central question posed in this paper is what are the causes of mobbing?
There have been heated discussions around this point. One view represented
by a section of the media and some employers, holds that mobbing victims are
responsible for being mobbed. Some physicians and clinical psychologists who
treat mobbing victims support the view. As reported by a number of case
studies, these physicians and psychologists cannot believe that the severe
symptoms observed in the victims are substantially a result of the victims'
work situation. Rather, they develop diagnoses such as ``querulous behavior'' or
The author would like to thank members of the Society against Psychosocial Stress and
Mobbing for their help in distributing questionnaires. The Konstanz sample was supported by a
fund of the University of Konstanz. The collection of the DAG-sample was organised and
International Journal of Manpower,
Vol. 20 No. 1/2, 1999, pp. 70-85.
funded by the Deutsche Angestellten Gewerkschaft Bezirk WuÈrttemberg. Thanks are also due
# MCB University Press, 0143-7720 to Ute Cornelius, Harald Krauss and Dieter Groeblinghoff.
``general anxiety disorder'' and suggest that these disorders existed before the Causes of
mobbing process began and that the disorders are what caused the mobbing bullying at work
process to develop.
Leymann (1996), as well as mobbing victims and their organisational
networks present a contrary view. This view claims that organisational
reasons, predominantly problems in the organisation of work and leadership
problems, are the cause of mobbing. 71
Yet another view can be found in the popular book by Adams (Adams with
Crawford, 1992) where certain individuals, the bullies, are made responsible for
workplace harassment. Finally, Schuster (1996), based on findings in
developmental and social psychology, points to the social system and processes
of social exclusion as a potential cause of mobbing.
There is insufficient research to date, to explain the causes of mobbing.
Methodological problems hamper the empirical investigation of the
phenomenon. As a result, there is an abundance of descriptive case reports that
are usually based on interviews with victims. The perspectives of perpetrators
and potential bystanders are not sought. In this paper, I will argue that one-
sided explanations should be avoided. Data will be presented to support the
view that multiple causes of mobbing have to be taken into consideration, and
that mobbing can be caused by more than one factor simultaneously. Figure 1
presents a model of the causes and consequences of mobbing. The four potential
causes of mobbing shown at the left-hand side of Figure 1 will be investigated:

Causes Mobbing Consequence

Organisational
Leadership
Organizational culture
Job stressors Rumours
Work organisation Psychosomatic
Social Isolation Complaints
Perpetrator Verbal Aggression Depression

Social Group Organisational Anxiety


Measures
Hostility Posttraumatic
Envy Attacking the Stress Disorder
Group pressure Private Sphere
Scapegoat Physical Obsession
Aggression
Attacking one’s
Person Attitudes
Personality
Qualification Figure 1.
Social skills Causes and
Stigmata
consequences of
mobbing
International the organisation, the mobbing perpetrator, the social system of the work group,
Journal of and the mobbing victim. In the case of a supervisor, there is usually some
Manpower overlap between the perpetrator, the social system and the organisation.
As can be seen in Figure 1, there are organisational and social factors, as
20,1/2 well as factors in the person of the mobber and of the victim which may cause
mobbing. Mobbing may lead to various forms of ill health. The figure also
72 illustrates one of the central problems in mobbing research: it is unclear
whether cause-effect relations go from the left to the right. There are, for
example, causes of mobbing in the organisation that may lead to observable
mobbing behavior. Mobbing may lead to health complaints in the victim.
However, one can equally assume that the causal path goes from the right to
the left. Anxious, depressive or obsessive behavior of the victim may produce a
negative reaction in the group, which leads to mobbing after some time (refer to
the literature showing that people tend to show negative responses to
depressive behaviour (Sacco et al., 1993)). Mobbing may lead to a decrease in
the social climate and social support, which in turn may weaken the
information flow. Again, this may increase job stressors such as uncertainty
(about goal achievement) and organisational problems.
What makes the discussion of the various potential causes of mobbing
difficult is that cause is often equated with guilt. For example, physically
disabled individuals may show a higher mobbing rate than the non-disabled. If,
to the best of our knowledge, all other possible explanations could be ruled out,
then even a cross-sectional study would not provide statistical evidence that
being physically disabled increases the likelihood of being mobbed (and not the
other way round). The reason is that variables such as gender, age or physical
handicaps cannot be influenced by other psychological variables (see Zapf et
al., 1996). However, it makes little sense to ``blame'' the physically disabled for
those variables. One might argue that if handicapped individuals carry a
higher risk of being mobbed in a certain organisation then this is a problem of
the social group, which is not able to deal with people who are different. That is,
the ``true cause'' in such cases may lie in the social group and not in the victim.
Obviously, the assignment of guilt is a question of interpretation.
Finally, leadership problems or organisational problems cannot ``harass'' an
employee. Such behaviour is only possible for human beings. In such cases,
there must always be people who react to these problems. In a statistical sense,
however, it is possible to speak of the organisation as the cause of mobbing. A
way to explain this is that mobbing is understood as an unresolved escalated
conflict, the frequency of which depends on the absolute frequency of conflicts in
a given organisation. If various organisational circumstances contribute to the
total number of conflicts, then the number of unresolved conflicts should also
increase, thus leading to a higher number of mobbing cases in the organisation.
The present exploratory study does not aim to ``prove'' any causal effects.
However, by combining several data sources and being mindful of the
objections that have been discussed, the paper provides empirical evidence for
all of the suggested causes of mobbing shown in Figure 1.
Method Causes of
Samples bullying at work
The research question was examined using two samples. The first sample
(Konstanz sample) was collected between October 1995 and July 1998, and
consisted of 96 mobbing victims and a control sample of N = 37. The mobbing
victims in this sample were recruited by means of newspaper articles on
mobbing, local broadcasting, mobbing self-help groups and by the help of a 73
German organisation called ``Society against Psycho-social Stress and Mobbing
GPSM''. The control group was collected by a snowball system starting with
persons personally known to members of the project. As far as possible, the
control group was matched with the mobbing group according to gender, age,
and education. The second sample (N = 118) was collected by the Deutsche
Angestellten-Gewerkschaft DAG Stuttgart. The DAG published an article
about the mobbing project in an internal journal, and interested job stewards
were asked to distribute questionnaires. Although the Konstanz sample
consisted of more severe mobbing cases (longer mean duration of mobbing,
higher scores on the mobbing scales below), the results of both samples were
very similar. For reasons of parsimony and clarity both samples were collapsed
for this article.
To define the mobbing group, two criteria that were used in the literature
(Zapf, 1999) were combined. First, in accordance with Leymann (1993b; 1996),
at least one out of a list of mobbing items had to be answered positively in the
sense that the specific mobbing behavior lasted at least six months and that
one of the mobbing behaviors occurred at least once a week. The second
criterion was that participants agreed that they were a victim of mobbing
according to the definition that was presented in the questionnaire. The
definition was ``Mobbing at work means harassing, bullying, offending,
socially excluding someone or assigning offending work tasks to someone in
the course of which the person confronted ends up in an inferior position''. The
definition was given after the list of mobbing items was presented (see
Einarsen and Skogstad, 1996, who used a very similar definition). To label
something mobbing (bullying) it has to occur repeatedly (for example, at least
once a week) and over a period of time (at least six months). It is not mobbing if
the incident is an isolated event or if two parties of approximately equal
``strength'' are in conflict. The mobbing group (N = 143) consisted of
participants for whom both criteria (item and definition) applied.
The second sample, the control group (N = 81), consisted of people for whom
both criteria did not apply. A total of 27 participants were excluded from the
study because of missing data or because only one criterion applied. Of the
sample, 57 percent were women. The mean age was 42 years and 30 percent of
the sample had some type of university degree.

Instruments
To measure potential causes of mobbing in the organisation, some scales of the
``Instrument for Stress-oriented Job Analysis'' (ISJA, version 5.1) (Semmer et al.,
International 1995; 1999; Zapf, 1993) were used. The scales were job complexity,
Journal of task-related job control, time-related job control, uncertainty, organisational
Manpower problems, concentration necessities, time pressure and compulsory co-
20,1/2 operation.
In addition, several scales measuring aspects of the social system were used.
A scale of social stressors (Frese and Zapf, 1987) comprised items referring to
74 the social climate in the work group, and conflicts with colleagues and
supervisors. A German version of the social support scales developed by
Caplan (Caplan et al., 1975) and adapted by Frese (1989) comprised, among
others, a subscale of social support by colleagues. Moreover, we translated the
Employee Opinion Survey of Lawthom et al. (1992), and a scale measuring a
positive communication climate was applied in this study.
Further, there were 45 items referring to aspects according to which the
mobbing victims saw themselves being different from the rest of the work
group (stigmatisation items, see Zapf and BuÈhler, 1998). From these items a
scale ``unassertiveness/avoidance'' was developed.
To measure conflict behaviour we used a German translation of the Rahim
Organisational Conflict Inventory ± ROCI II (Rahim and Magner, 1995)
measuring five conflict styles: avoiding, compromising, dominating,
integrating and obliging.
Mobbing behaviour was measured with the help of a German translation of
the Leymann Inventory of Psychological Terrorisation LIPT (Leymann, 1990).
We extended the Leymann questionnaire and constructed six mobbing scales
published in Zapf et al. (1996). The first scale was mobbing by organisational
measures, which included questioning a person's decisions, judging a person's
job performance wrongly or in an offending manner, and assigning degrading
tasks to the person concerned. The second scale was social isolation, which
included one does not talk to the person concerned, and being treated like air or
non-existent. The third scale was attacking the private sphere which included
permanently criticizing a person's private life, making a person look stupid,
and suspecting a person to be psychologically disturbed. The fourth scale was
verbal aggression, which included shouting at or cursing loudly at a person,
and verbal threats. The fifth scale was physical aggression, which included
threat of physical violence, and minor use of violence. The sixth scale was
rumours, which included saying nasty things about a person behind his or her
back.
Psychological dysfunctioning was measured using various scales developed
by Mohr (1986). Self-esteem was measured with items like ``I am proud of
my achievements''. Depression was measured with items like ``I feel lonesome,
even when together with other people'' or ``I look forward to the future
hopelessly''. Anxiety comprised items such as ``I prefer to avoid difficult
situations'' or ``Sudden noise at night frightens me''. A scale of psychosomatic
complaints included items such as ``how often do you suffer from headaches,
high blood pressure, insomnia''. In addition, we applied a German translation of Causes of
the PANAS scale of Watson et al. (1988) measuring negative and positive bullying at work
affect.
Finally, we used items first developed by Knorz to ask the participants of the
study for their opinion about why they experienced mobbing (see Zapf et al.,
1995; see the description below). Psychometric properties of the scales are
summarised in Table I. 75

Cronbach's Number of Range of


alpha Mean SD items scale

Causes of mobbing:
the organisation
Job complexity 0.89 3.21 0.83 10 1-5
Job control/tasks 0.90 3.27 0.98 7 1-5
Job control/time 0.87 3.32 1.13 5 1-5
Uncertainty 0.77 2.66 0.85 5 1-5
Organisational problems 0.73 2.96 0.65 8 1-5
Concentration necessities 0.73 3.12 0.81 5 1-5
Time pressure 0.87 3.14 0.94 6 1-5
Compulsory cooperation 0.82 2.80 0.93 6 1-5
The social system
Social stressors 0.85 2.31 0.74 8 1-4
Social support by
colleagues 0.92 2.41 0.83 4 1-4
Communication climate 0.90 2.32 0.52 8 1-5
The victim
Unassertiveness/
avoidance 0.71 1.28 0.26 7 1-5
Avoiding 0.85 3.08 0.93 6 1-5
Compromising 0.62 2.24 0.79 3 1-5
Dominating 0.76 2.78 0.89 4 1-5
Integrating 0.86 3.68 0.81 7 1-5
Obliging 0.80 3.44 0.69 6 1-5
Mobbing behaviour
Organisational measures 0.92 2.33 1.07 13 1-5
Social isolation 0.81 2.07 1.06 4 1-5
Attacking the private 0.74 1.82 0.94 5 1-5
sphere
Verbal aggression 0.85 2.62 1.15 6 1-5
Physical aggression 0.54 1.24 0.57 4 1-5
Rumours 0.70 3.32 1.82 2 1-5
Consequences of mobbing:
the victim's wellbeing
Self-esteem 0.68 4.29 0.49 8 1-5
Depression 0.86 3.04 1.04 8 1-7
Anxiety 0.64 2.40 0.84 7 1-6
Psychosomatic
complaints 0.93 2.70 0.89 20 1-5 Table I.
Negative affect 0.86 1.69 0.69 10 1-5 Descriptive data of the
Positive affect 0.87 3.20 0.72 9 1-5 variables
International Results
Journal of The perpetrator as a cause of mobbing
Manpower First, we analysed a series of questions about the reasons for mobbing from the
victims' perspective. The first item listed was: ``Because of which reasons have
20,1/2
you been treated so negatively by your colleagues, supervisors or
subordinates?'' As can be seen in Table II, the most frequently reported reason
76 was ``They wanted to push me out of the company''. Typically, this is believed
to be the final goal of a mobbing process (Leymann, 1993b; 1996). Only 9
percent disagreed with this statement. The second most frequently reported
reason was ``a hostile person influenced others''. At least from the victims'
perspective the cause of mobbing is often seen in a specific perpetrator. These
results replicate findings of an earlier study (Zapf et al., 1995).

The organisation as a cause of mobbing


Table II shows that the organisational climate, high stress and organisational
problems are among the most frequent causes for mobbing from the victims'
perspective. Two subsequent analyses were conducted on these results. First,
we compared mobbing victims and the control group with regard to working

Causes Yes Partly No

They wanted to push me out of the company 66 25 9


A hostile person influenced others 62 16 22
Organisational climate 49 37 14
High stress 45 29 26
Unresolved conflict 41 33 26
They didn't like me 32 40 28
Organisational problems 32 30 38
Envya 30 33 37
I had a privileged position in the firm 30 26 45
Did not adapt to the rules of the majority 29 34 37
Performance above average 27 34 39
Because I am different 21 16 63
Illness 15 10 75
Love of mockery 13 34 53
Because everybody is treated alike 12 37 51
Boredom 10 17 73
Because of my private life 8 24 68
Because of religious/political attitudes 8 14 78
Because of my gender 8 9 83
Because of a bodily handicap 6 6 87
Because of my appearance 5 14 81
Performance below average 2 11 87
Because of my nationality 2 4 94

Table II. Notes:


Causes of mobbing: the N
a
= 123-136
victim's view (in %) N = 78
conditions (Zapf and Osterwalder, 1998). Clearly, the mobbing group differed Causes of
from the control group with regard to all job characteristics. There were higher bullying at work
stressors and less job control for the mobbing group. Further, the strongest
effects appeared for task and time-related control, uncertainty and
organisational problems. Interestingly, job complexity was the only item where
no differences appeared. This probably reflects the fact that mobbing can occur
at lower and higher levels of the organisational hierarchy. Typically, at the 77
lower levels, there are also lower levels of occupational skills, and less complex
tasks. While at higher levels of the organisation skill levels are higher,
managers and specialist workers who work on more complex tasks also
become the target of workplace harassment.
To validate the self-reported statements of mobbing victims about the
causes of mobbing, we chose the item ``organisational problems'' (see Table II).
An analysis of variance was conducted by comparing four groups: individuals
disagreeing, partly agreeing, and agreeing with the above mentioned item, plus
the control group. Significant results were found when ``organisational
problems'' from the job analysis instrument (see Table I) were used as the
dependent variable. The Scheffe post hoc test (p < 0.05) showed that the control
group differed significantly from all other groups. However, the group
disagreeing with the statement that organisational problems are a cause, also
reported significantly less organisational problems than those who agreed with
this item.
These analyses support the victim's interpretation that the organisation is the
reason why mobbing occurred. It is unlikely that this result is a mere artifact of
the self-report data because the information was gathered in quite different
ways. The participants had to answer the job analysis questionnaire first. It
would be impossible for them to assess whether they are high or low compared
to others, and so give the ``correct'' answer when they responded to the ``causes of
mobbing'' questions. While bad working conditions can result from the mobbing
process, it seems unlikely that this holds for all cases. In combination with the
``reasons for mobbing'' items it is claimed that organisational problems are the
cause of mobbing in at least some cases (see Zapf and Osterwalder, 1998).

The victim and the social system as causes of mobbing


In the list of reasons in Table II, there are several items suggesting that the
reasons for at least some of the mobbing cases may lie in the victims
themselves. It is interesting to see that these items are mostly at the end of the
table. Altogether, 32 percent of the victims agree with one of these items. As
discussed in the introduction, there is a difference between finding a cause and
assigning guilt or responsibility. It is sometimes difficult to say whether the
cause is in the victim or in the social group that, for some reason, is not able to
integrate a person who is different from them in some respect.
It is noticeable that only 2 percent of the victims ``admitted'' that their
performance was below average. Practical experience suggests that this
figure should be higher, and the question arises as to whether mobbing victims
International are open to taking such personal reasons into account. Examples of
Journal of personal reasons include deficits in social skills, low performance, ``being
Manpower difficult'' (such as being pedantic about accuracy), or being aggressive or
moaning.
20,1/2 The list of 45 items referring to aspects according to which the mobbing
victims saw themselves being different from the rest of the work group (Zapf
78 and BuÈhler, 1998) comprised several items referring to a lack of social skills and
unassertive behaviour. An example is ``I do not recognise conflicts as quickly as
others''. Zapf and BuÈhler (1998) constructed a scale ``Unassertiveness/
avoidance'' and investigated the relationship with other variables. A
hypothesis was that being unassertive, tending to avoid conflicts, or showing
little effort to be a member of the group increase the likelihood of becoming a
mobbing victim. Clearly, not all mobbing victims would have low levels of
communication or social skills. Rather, it is likely that there is a subgroup that
have lower social skills levels, while others may not differ from the control
group.
Some of the items listed as ``reasons for mobbing'' supported the hypothesis
about the victim as a cause. Subjects high in unassertiveness/avoidance more
often said that their performance was below average (2 = 7.95, p = 0.093); that
they were bullied because of their nationality (2 = 9.66, p = 0.047); and
because of a bodily handicap (2 = 8.70, p = 0.069). Results for conflict style
and psychological wellbeing are shown in Table III. People high in

Unassertiveness/avoidance
Control
group Low Medium High Overall significance
1 2 3 4 and significantly
N = 84 N = 29 N = 57 N = 57 differing groups

Conflict style
Avoiding 2.80 2.60 3.17 3.54 ** 1, 2 vs 4; 2 vs 3
Compromising 3.39 3.54 3.48 3.06 * 2, 3 vs 4
Dominating 2.86 2.78 2.77 2.60 n.s.
Integrating 3.77 3.94 3.75 3.42 * 2 vs 4
Obliging 3.38 3.34 3.37 3.65 p = 0.07
Well-being
Self-esteem 4.31 4.39 4.34 4.16 n.s.
Anxiety 2.19 2.12 2.32 2.90 ** 1, 2, 3 vs 4
Depression 2.61 2.94 3.08 3.73 ** 1, 2, 3 vs 4; 1 vs 3
Psychosom. comp. 2.10 2.80 3.12 3.11 ** 1 vs 3, 4
Table III. Negative affect 1.41 1.57 1.71 1.99 ** 1,2 vs 4; 1 vs 3
Conflict styles and Positive affect 3.17 3.28 3.29 3.12 n.s.
psychological wellbeing
for mobbing victims Notes:
with different levels of ** p < 0.01
unassertiveness/ * p < 0.05
avoidance and control n.s.: not significant
group Post hoc tests Scheffe p < 0.10
unassertiveness/avoidance showed the worst conflict resolution behaviour Causes of
with regard to all strategies except dominating. In three of five cases, this bullying at work
group differed from the low unassertiveness/avoidance group. The strongest
effect appeared for the ``avoiding'' conflict style. This is not surprising given the
high content overlap between the unassertiveness/avoidance scale and the
ROCI avoidance scale. Moreover, the low unassertiveness/avoidance group
showed results comparable with those of the control group. 79
The group high in unassertiveness/avoidance reported a significantly higher
level of anxiety and depression compared with all other groups and a higher
level of negative affect than the control group and the group low in
unassertiveness. By contrast, psychosomatic complaints, which are a typical
strain indicator used in organisational stress research, differentiated the control
group from all mobbing groups. It is interesting to note that concepts such as
anxiety, depression and negative affect have been related to neuroticism.
Watson and Clark (1984) differentiated the high unassertiveness/avoidance
group from all other groups, whereas the strain indicator psychosomatic
complaints differentiated the groups that are under high stress, namely the
mobbing groups from the control group. It is not suggested here that the
variables related to neuroticism cannot be affected by environmental
circumstances at all (see the discussion in Spector et al., in press). However, in
the present context, the trait-like character of neuroticism variables has to be
taken into consideration. Based on these findings, it can be hypothesised that
there is a group of individuals who had pre-existing symptoms of anxiety,
depression and negative affect. This group of mobbing victims has lower social
skills than their colleagues and show deficiencies in their social behaviour and,
thus, they have an increased likelihood of becoming a victim of mobbing. These
individuals are not very sensitive to conflict, and they are unable to build up a
stable social network. They avoid conflict as far as possible, and, if this is
impossible, they tend to give way. By contrast, there are other victims of
mobbing who do not show these characteristics and who cannot be statistically
differentiated from the control group with the exception of psychosomatic
complaints. The latter result indicates that regardless of the cause, mobbing is
a severe stressor.
As a final step, we tried to build groups according to the four mentioned
causes of mobbing using all the information described in this article (see Table
IV). The ``organisation as cause group'' consisted of those agreeing that
organisational problems were a cause of mobbing (see Table II) and showed a
high score on the organisational problems scale (> 3.0), but who did not say
that that their social climate was extremely harsh. Members of this group were
also not in the extreme group of unassertiveness/avoidance. The ``perpetrator
group'' agreed with the item that there was one hostile person in the
organisation (see Table II). However, the ``perpetrator group'' did not fulfil the
criteria of the other groups, such as scores on the organisational problems
scale. The ``social system group'' showed low scores on the social support by
colleagues scale (< 2.0) and the scale measuring the communication climate
International Social Overall significance
Journal of Organisation Perpetrator system Victim and significantly
1 2 3 4 differing groups
Manpower N = 24 N = 36 N = 22 N = 35 (groups = 1, 2, 3, 4)
20,1/2
The organisation
Job complexity 3.55 3.06 3.33 3.14 n.s.
80 Job control/tasks 3.09 2.94 2.92 3.10 n.s.
Job control/time 3.22 3.33 2.85 2.79 n.s.
Uncertainty 3.22 2.51 3.15 2.65 ** 1, 3 vs 2; 1 vs 4
Organisation
problems 3.46 3.00 3.46 2.98 * 1, 3 vs 2, 4
Concentration
necessary 3.43 3.03 3.43 3.03 p = 0.09
Time pressure 3.22 3.14 3.82 3.00 * 3 vs 2, 4
Compulsory
cooperation 2.84 2.65 2.89 2.77 n.s.
The social system
Social stressors 2.78 2.51 3.21 2.60 ** 3 vs 2, 4
The victim
Avoiding 3.43 2.83 2.88 3.48 ** 2 vs 4
Compromising 3.39 3.28 3.14 3.07 n.s.
Dominating 2.97 2.71 2.38 2.63 n.s.
Integrating 3.50 3.85 3.83 3.36 * 2 vs 4
Obliging 3.36 3.51 3.18 3.67 n.s.
Mobbing behaviour
Organisation 2.91 2.62 3.33 2.96 p = 0.06 2 vs 3
measures
Social isolation 2.51 2.53 3.11 2.65 n.s.
Attack private
sphere 2.05 1.97 2.82 2.28 ** 1, 2 vs 3
Physical aggression 1.17 1.28 1.53 1.26 n.s.
Verbal aggression 3.14 3.15 3.74 3.34 p = 0.08
Rumours 3.75 4.26 4.55 4.53 n.s.
Wellbeing
Self-esteem 4.30 4.41 4.41 4.08 * 2 vs 4
Anxiety 2.50 2.38 2.18 2.84 * 3 vs 4
Depression 3.33 2.97 2.97 3.82 ** 2, 3 vs 4
Psychosomatic
comp. 3.11 2.88 3.29 3.17 n.s.
Negative affect 1.79 1.59 1.74 2.08 * 2 vs 4
Table IV. Positive affect 3.13 3.26 3.65 3.08 p = 0.06 3 vs 4
Causes of mobbing and
their relationship with Notes:
job characteristics, ANOVA
mobbing behavior, ** p < 0.01 * p < 0.05
conflict styles and n.s.: not significant
wellbeing Post hoc tests Scheffe p < 0.10

(< 2.0). In addition, they agreed with the item that the organisational climate
was a reason for mobbing (Table II). Finally, the ``victims group'' was identical
with the group high in unassertiveness/avoidance in Table III.
Looking first at job characteristics shows that the ``social system as a cause of Causes of
mobbing'' group and the ``organisation as a cause of mobbing'' group showed bullying at work
worse working conditions than the others. Significant results appeared for
uncertainty, organisational problems and time pressure. Not surprisingly, the
``social system'' group showed the highest social stressors at work.
With regard to conflict styles, the ``victims'' group differed from the
``perpetrator'' group. Significant results appeared for the avoiding and 81
integrating conflict styles. With regard to mobbing, the ``social system as a
cause'' group had to deal with the highest level of mobbing and the
``perpetrator'' group with the lowest level. Interestingly, significant differences
appeared only in relation to the item ``attacking the private sphere'', and a
tendency appeared for the items ``organisational measures'' and ``verbal
aggression''.
With regard to wellbeing, the ``victims'' group showed higher scores for
anxiety, depression and negative affect than the other groups, and lower scores
in self-esteem. Comparatively low levels occurred for the ``social system'' and
the ``perpetrator'' groups. However, no such differences occurred for
psychosomatic complaints. Although it is not statistically significant, it is the
``social system'' group that reports the highest level of psychosomatic
complaints.

Discussion
This article investigated causes of mobbing at work. Being aware that it is
difficult to prove cause and effect with regard to mobbing at the current stage
of mobbing research, material was presented suggesting that various causes of
mobbing have to be considered rather than favoring one-sided explanations.
Unfortunately, such explanations dominate some of the scientific and popular
literature on mobbing, and among practitioners. Some physicians prefer the
view that the mobbing victims are the cause of mobbing. Employers sometimes
tend to share this position and from many anecdotal reports it is known that
the victims, not the perpetrators, are forced to leave the organisation (Leymann,
1993b; 1996). From the perspective of the organisation this is often easier,
particularly if managers are among the mobbers. Physicians may often be
unaware that they deal with a selected group of mobbing victims. Those
individuals who are more or less accidental victims of a bully usually have no
long-term contact with physicians. These victims are often surprised at
suddenly being the victim of harassment after a long period of working life
without any substantial problems and they desperately try to solve the conflict,
often by leaving the organisation and finding another job elsewhere. For these
victims, there is usually no need for a psychiatric assessment. Psychiatrists are
more often confronted with individuals who belong to the ``victims as a cause''
group. As shown in Table IV, these victims suffer from more serious health
symptoms than the other groups. They are more often in need of psychiatric
treatment, and they are frequently unable to continue working while being
mobbed. These victims are less able to recover from the mobbing process once
International it is over, and they are more often engaged in fighting their case long after
Journal of leaving the organisation.
Manpower It is interesting to see that a single perpetrator seems to play a more
important role in the British literature (Adams with Crawford, 1992; Field,
20,1/2 1996; Rayner, 1997). As mentioned in the introduction, this may partly be
due to the terminology used. The Scandinavian and German research seems to
82 be more focussed on the victim attacked by one or more persons, whereas
British research and practice seems to be more focussed on the attacker. It is
difficult to assess whether the finding (see Table II) showing 62 percent of
mobbing to be caused by a certain perpetrator is realistic. There may be cases
where the cause lies in the organisation but is manifested by a specific
individual. Alternatively, the cause could lie primarily in the social system,
with a specific individual seen as the ringleader. Attribution theory (Baron,
1990) suggests that people tend to make personal attributions. That is, they
prefer to blame other people for an action rather than attribute the situation to
non-human factors such as time pressures or job complexity. However, even if
the above mentioned 62 percent is an exaggerated figure, it seems unlikely that
almost two-thirds of the sample completely misperceive or misinterpret their
situation.
Leymann (1993b) reports many cases showing that poor work organisation
and leadership problems are related to mobbing. Results from Einarsen et al.
(1994), Vartia (1996) and Zapf and Osterwalder (1998) support this view.
They all found relationships between the frequency of mobbing and
organisational variables. However, as shown in Figure 1, it must be considered
that negative working conditions can be a result of mobbing. This may arise
from a lack of communication, conscious miscommunication, which in itself is a
mobbing strategy, or various kinds of conflicts that affect cooperation and
information flow. On the other hand it seems plausible that in workplaces that
are low in control and high in organisational problems, for example, time
pressure, uncertainty due to unclear responsibilities and role ambiguities, the
probability of conflict due to these organisational problems is high. The idea is
simple: a higher base rate of conflict in organisations leads to a higher rate of
escalating conflict which in turn are the basis for mobbing. Moreover, high
scores in organisational problems and uncertainty might also be indications of
long-term, entrenched conflict that heighten the potential for mobbing to
develop.
It is interesting to see that the ``victims'' group, which is rather low in work
and social stressors, and medium in mobbing, shows the highest scores in
anxiety, depression and negative affect. By contrast, the ``social system'' group,
which is highest in job stressors, social stressors and mobbing, show the lowest
scores in anxiety, depression and negative affect. While it is not statistically
significant, the ``social system'' group shows the highest mean in psychosomatic
complaints. Anxiety and depression have been discussed both as independent
and dependent variables in the stress process (e.g., Spector et al., in press).
Whereas psychosomatic complaints are somewhat higher in the group where
work and social stressors and mobbing is high, and thus reflect the idea that Causes of
stressors cause strain, this is not the case for anxiety and depression. This bullying at work
picture can, therefore, be taken as an indication that a number of factors
contribute to the likelihood of becoming a victim of mobbing. Such factors
include a high level of anxiety and depression and the related behaviours of
conflict avoidance; inability to recognise conflict; being shy; and showing little
effort to integrate in the work group. On the other hand, it is clear from the data 83
presented that not all victims of mobbing are low in social competencies. As
shown in Table III, the mobbing group low in unassertiveness/avoidance does
not differ from the control group with regard to conflict behaviour and
wellbeing. The exception is psychosomatic complaints.
The literature on communication and conflict has always pointed to the
difficulties in analysing the causes of conflicts (Glasl, 1994; Thomas, 1992).
First, there is the problem of where to stop in the causal chain of conflict
development. Second, there is often more than one factor that contributes to the
escalation of a conflict. There may be a person with a tendency to harass others
(i.e. the perpetrator as a cause for mobbing). For the perpetrator, it may be
easier to harass a person who is an out-group member because of his or her
inability to integrate into the group (victim characteristics as causes of
mobbing). It may be even easier to harass someone if there are working
conditions with high uncertainty and high rates of organisational problems.
Such working conditions may lead an individual to commit multiple errors at
work, which, in turn, can be used as ammunition against the individual (the
organisation as a cause of mobbing). The occurrence of mobbing may be
further supported if there are tensions in the work group for which a potential
scapegoat is a good `` lightning conductor'' (the social system as a cause of
mobbing). Moreover, there may be factors that contribute to the emergence of
conflict, and other factors that contribute to its escalation.
Finally, it has already been mentioned that the identification of a cause of
mobbing is sometimes a question of interpretation. If a physical handicap is
taken as a reason to harass a person, one may argue that the cause of mobbing
lies in the social group that is unable to deal with people who are different.
There may be more interpretation problems in the case of a high performing
person who is bullied ± high achievers are a typical group of mobbing victims
(see Zapf and BuÈhler, 1998). Partly, such problems might occur because it may
be unclear whether the high performing person does not conform to group
norms, behaviour that is not tolerated by the rest of the group. Alternatively, a
group cannot tolerate the high performing person who demonstrates his or her
achievements in an arrogant and provocative manner.
The central argument in this article is that there are potentially multiple
causes of mobbing. The results presented here are based on an empirical study
using a group of mobbing victims and a control sample. Clearly, further
empirical research is needed to increase our understanding of the causes of
mobbing and implications for individuals, organisations, and society.
International References
Journal of Adams, A. with Crawford, N. (1992), Bullying at Work, Virago Press, London.
Manpower Baron, R.A. (1990), ``Attributions and organizational conflict'', in Graham, S. and Volkes, V. (Eds),
Attribution Theory: Applications to Achievement, Mental Health and Interpersonal Conflict,
20,1/2 Erlbaum, Hillsdale, NJ, pp. 211-30.
Caplan, R.D., Cobb, S., French, J.R.P., van Harrison, R. and Pinneau, S.R. (1975), Job Demands and
84 Worker Health, National Institute for Occupational Safety and Health (NIOSH),
Washington, DC.
Einarsen, S. and Skogstad, A. (1996), ``Prevalence and risk groups of bullying and harassment at
work'', European Journal of Work and Organisational Psychology, Vol. 5, pp. 185-202.
Einarsen, S., Raknes, B.I. and Matthiesen, S.B. (1994), ``Bullying and harassment at work and
their relationships to work environment quality: an exploratory study'', The European
Work and Organisational Psychologist, Vol. 4, pp. 381-401.
Field, T. (1996), Bullying in Sight, Success Unlimited.
Frese, M. (1989), ``GuÈtekriterien der Operationalisierung von sozialer UnterstuÈtzung am
Arbeitsplatz'', (``Psychometric criteria for the operationalisation of social support at work''),
Zeitschrift fuÈr Arbeitswissenschaft, Vol. 43, pp. 112-22.
Frese, M. and Zapf, D. (1987), ``Eine Skala zur Erfassung von Sozialen Stressoren am
Arbeitsplatz'', (``A scale for the assessment of social stressors at work''), Zeitschrift fuÈr
Arbeitswissenschaft, Vol. 41, pp. 134-41.
Glasl, F. (1994), Konfliktmanagement. Ein Handbuch fuÈr FuÈhrungskraÈfte und Berater, (Conflict
Management. A Handbook for Managers and Consultants), 4th ed.), Haupt, Bern.
Lawthom, R., Maitlis, S., Patterson, M. and West, M.A. (1992), ``Employee opinion survey'',
unpublished questionnaire, Medical Research Council.
Leymann, H. (1990), Presentation av LIPT-formulaÈret: Konstruktion, validering, utfall,
(Presentation of the LIPT Questionnaire: Construction, Validation And Outcome), Violen
inom PraktikertjaÈnst, Stockholm.
Leymann, H. (1993a), ``A È tiologie und HaÈufigkeit von Mobbing am Arbeitsplatz ± eine U È bersicht
uÈber die bisherige Forschung'', (``Etiology and frequency of mobbing at work ± a review of
existing research''), Zeitschrift fuÈr Personalforschung, Vol. 7, pp. 271-83.
Leymann, H. (1993b), Mobbing ± Psychoterror am Arbeitsplatz und wie man sich dagegen wehren
kann, (Mobbing ± Psychological Terror at Work, and How One Can Defend Oneself),
Rowohlt, Reinbeck.
Leymann, H. (1996), ``The content and development of mobbing at work'', European Journal of
Work and Organisational Psychology, Vol. 5, pp. 165-84.
Mohr, G. (1986), Die Erfassung psychischer BefindensbeeintraÈchtigungen bei Arbeitern, (The
Measurement of Psychological Dysfunctioning of Workers), Peter Lang, Frankfurt a.M.
Rahim, M.A. and Magner, N.R. (1995), ``Confirmatory factor analysis of the styles of handling
interpersonal conflict: first-order factor model and its invariance across groups'', Journal of
Applied Psychology, Vol. 80, pp. 122-32.
Rayner, C. (1997), ``The incidence of workplace bullying'', Journal of Community and Applied
Social Psychology, Vol. 7, pp. 199-208.
Sacco, W.P., Dumont, C.P. and Dow, M.G. (1993),``Attributional, perceptual, and affective
responses to depressed and nondepressed marital partners'', Journal of Consulting and
Clinical Psychology, Vol. 61, pp. 1076-82.
Schuster, B. (1996), ``Rejection, exclusion, and harassment at work and in schools'', European
Psychologist, Vol. 1, pp. 293-309.
Semmer, N.K., Zapf, D. and Dunckel, H. (1995), ``Assessing stress at work: a framework and an Causes of
instrument'', in Svane, O. and Johansen, C. (Eds), Work and Health ± Scientific Basis of
Progress in the Working Environment, Office for Official Publications of the European bullying at work
Communities, Luxembourg, pp. 105-13.
Semmer, N.K., Zapf, D. and Dunckel, H. (1999), ``Stress-orientierte Arbeitsanalyse ISTA'', ``Stress-
oriented job-analysis ISTA'', in Dunckel, H. (Ed.), Handbuch zur Arbeitsanalyse, Verlag der
Fachvereine, ZuÈrich.
Spector, P.E., Zapf, D., Chen, P.Y. and Freese, M. (in press), ``Why negative affectivity should not 85
be controlled in job stress research: don't throw the baby out with the bath water'', Journal
of Organizational Behaviour.
Thomas, K.W. (1992),``Conflict and negotiation processes in organisations'', in Dunnette, M.D.
and Hough, L.M. (Eds), Handbook of Industrial and Organisational Psychology, Vol. 3,
Consulting Psychologists Press, Palo Alto, CA, pp. 651-718.
Vartia, M. (1996), ``The sources of bullying ± psychological work environment and organisational
climate'', The European Journal of Work and Organisational Psychology, Vol. 5, pp. 203-14.
Watson, D. and Clark, L.A. (1984), ``Negative affectivity: the disposition to experience aversive
emotional states'', Psychological Bulletin, Vol. 96, pp. 465-90.
Watson, D., Clark, L.A. and Tellegen, A. (1988), ``Development and validation of brief measures of
positive and negative affect: the PANAS scales'', Journal of Personality and Social
Psychology, Vol. 54, pp. 1063-70.
Zapf, D. (1993), ``Stress-oriented job analysis of computerised office work'', The European Work
and Organisational Psychologist, Vol. 3, pp. 85-100.
Zapf, D. (1999), ``Mobbing in Organisationen. Ein U È berblick zum Stand der Forschung'',
(``Mobbing in organisations. A state of the art research review''), Zeitschrift fuÈr Arbeits- &
Organisationspsychologie, Vol. 43. pp. 1-25.
Zapf, D. and BuÈhler, K. (1998), Exclusion and Stigmatisation at Work, Department of Psychology,
J.W. Goethe-University Frankfurt.
Zapf, D. and Osterwalder, P. (1998), Organisational Causes of Workplace Harassment,
Department of Psychology, J. W. Goethe-University Frankfurt.
Zapf, D., Dormann, C. and Frese, M. (1996), ``Longitudinal studies in organisational stress
research: a review of the literature with reference to methodological issues'', Journal of
Occupational Health Psychology, Vol. 1, pp. 145-69.
Zapf, D., Knorz, C. and Kulla, M. (1995), ``Causes, coping, and consequences of various mobbing
factors at work'', paper presented at the 7th European Congress of Work and
Organisational Psychology, 19-22 April, GyoÈr, Hungary.
Zapf, D., Knorz, C. and Kulla, M. (1996), ``On the relationship between mobbing factors, and job
content, the social work environment and health outcomes'', European Journal of Work
and Organisational Psychology, Vol. 5, pp. 215-37.

You might also like