phd thesis proposalprr.hec.gov.pk/jspui/bitstream/123456789/6866/1/phd... · the current thesis...
TRANSCRIPT
1
WORKPLACE BULLYING AND ITS RELATIONSHIP WITH
EMPLOYEES’ JOB STRAINS: THE ROLE OF INDIVIDUAL AND
AFFECTIVE FACTORS
PhD Thesis Proposal
Researcher:
Saima Naseer 24-FMS/PHD-MGT/F10
Supervisor:
Dr. Mohammad Bashir Khan
Faculty of Management Sciences
INTERNATIONAL ISLAMIC UNIVERSITY,
ISLAMABAD
2
WORKPLACE BULLYING AND ITS RELATIONSHIP WITH
EMPLOYEES’ JOB STRAINS: THE ROLE OF INDIVIDUAL AND
AFFECTIVE FACTORS
(A Doctoral Dissertation Proposal PhD-Management)
Saima Naseer
Reg # 24-FMS/PHDMGT/F10
Submitted in partial fulfillment of the requirements for the
PhD Degree with the specialization in Management
at the Faculty of Management Sciences,
International Islamic University,
Islamabad.
3
CHAPTER 1
INTRODUCTION
1.1 Background
In the current era of globalization, the most pervasive problem faced by nearly every
organization is the emerging, prevalent yet uncontrolled phenomenon of workplace bullying
(Hutchinson, Vickers, Jackson & Wilkes, 2010). Workplace bullying in the past decade has been
recognized as one of the important characteristic of the workplace which has generated rising
interest and attention from not only researchers but practitioners as well (Einarsen, Raknes, &
Matthiesen, 1994; Hutchinson et. al., 2010; Leymann, 1990a; Samnani & Singh, 2012). The
persistent nature of bullying is acknowledged to such an extent that researchers admit the fact
that this phenomenon is an inescapable and harsh reality of today’s workplace which is shifting
from the schools to the offices and boardrooms of the workplace (Harvey, Heames, Richey, &
Leonard, 2006).
The concept of workplace bullying has been examined by researchers through different labels for
instance Workplace Harassment (Bjorkqvist, Osterman & Hjelt-Back, 1994), Emotional abuse
(Keashly, 1998), workplace victimization (Einarsen & Raknes, 1997), Mobbing (Leyman 1996),
psychological terror (Leymann, 1990a) and bullying at the workplace (Hoel, Cooper & Faragher,
2001).
Although a number of terminologies exist, however presently there is agreement over researchers
on what behaviors constitute bullying at the workplace. Workplace bullying refers to irritating,
4
intimidating, socially boycotting and creating obstacles in one’s work related tasks (Einarsen,
2000; Einarsen, Hoel, Zapf & Copper, 2011). For bullying to occur there is an agreement among
researchers that bullying behaviors are those that occur frequently and continuously and they
range in duration (Einarsen et al., 2011). Moreover, bullying is an aggravating process which
tends to make the victim end up in a situation where he/she is unable to defend himself/herself
and become the subject of continuous harmful behaviors (Salin, 2003; Zapf, Einarsen, Hoel &
Vartia, 2003). A disagreement or clash is not considered bullying if the event is a one time event
or if the individuals involved in the clash hold equal power and authority (Einarsen, Hoel, Zapf,
& Cooper, 2003).
The rising attention in the workplace bullying area has resulted in a substantial literature that has
investigated the types of workplace bullying behaviors (Einarsen et.al., 2003); the occurrence of
such abusive behaviors at the workplace (Zapf et.al., 2003); the harmful consequences of
bullying on the targets and workplace (Giradi, Monaco, Prestigiacomo, Talamo, Ruberto, &
Tatarelli, 2007); and avoidance and handling processes pertaining to workplace bullying
behaviors (Saam, in press; Schat & Kelloway, 2003). While past studies on bullying have
devoted a substantial consideration in investigating occurrence, determinants and outcomes
(Bowling & Beehr, 2006; Hershcovis, Turner, Barling, Inness, et.al., 2007; Zapf & Einarsen,
2005), research underlying the various individual mechanisms through which bullying shows its
deleterious effects is still lacking (Cassidy, McLaughlin & McDowell, 2014; Nielsen &
Einarsen, 2012; Samnani & Singh, 2012; Tuckey & Neall, 2014).
According to the stressor-strain model, workplace bullying has been regarded as an important
stressor which causes strain within organizations and is one of the most distressing issues in
5
today’s workplace as in comparison to other organizational stressors (Einarsen & Mikkelsen,
2003; Zapf, Knorz & Khulla, 1996b).
However, researchers contend that conceptualizing bullying simply as a stressor-strain model is
too simplistic and there might be important mediators (Coyne, Clough, Alexander, & Clemment,
2006; Mikkelsen & Einarsen, 2002) existing in the workplace bullying and outcomes
relationship which might further our understanding the process of workplace bullying.
In this context, a few studies in the past have suggested how, why and under what conditions a
stressor may predict strain over time (Zapf, Dormann, & Frese, 1996a). These theorists suggest
that the impact of stressor with the passage of time and through certain mechanisms have the
capacity to increase strain over time.
Empirically this argument has been tested in the bullying context where high levels of exposure
to workplace bullying may decrease an individual’s sense of vigor and dedication which might
lead to increase strain with the passage of time (Munoz, Baillien, Witte, Jimenez & Pastor,
2009). In other words, work engagement has been found as a mediator between workplace
bullying incidences and strain. Moreover, lower social support has been found as a partial
mediator between workplace bullying and anxiety reactions (Hansen, Hogh, Persson, Karlson,
Garde & Orbeak, 2006). Giorgi (2012) in a survey of 300 Italian academic staff employed in a
large sized Italian university reported that bullying at the workplace resulted in harmful health
outcomes through its effects on workplace climate. Their findings lend credence to the fact that
workplace bullying causes negative organizational climate which ultimately hampers one’s
health (Giorgi, 2012).
6
Recently, Tuckey & Neall (2014) in a between and within person examination using multisource
data and two study longitudinal research study found that emotional exhaustion partially
mediated the workplace bullying and optimism and bullying-self efficacy relationship. These
authors call for future researchers to further explore and study additional affective, cognitive and
personal resources as mechanisms through which workplace bullying shows its negative effects
on outcomes.
Thus, there is a rising outcry by researchers that the workplace bullying is a complex process
(Monks, Smith, Naylor, Barter, Ireland & Coyne, 2009) instead of viewing it simply as a
stressor-strain model, researchers need to explain and interpret bullying by taking into account a
number of intervening processes between the bullying and outcomes relationship. In addition,
various theoretical models need to be incorporated in this area to effectively understand the
mechanisms through which work related bullying emerges (Hoel et.al., 2003).
The appraisal theory of stress has been acknowledged as an authority in psychology (Lazarus &
Folkman, 1984). This theory considers cognitions and emotions as mediators between stressor-
strain relationships. Moreover, conservation of resource theory (Hobfoll, 1998; 2001) an
extension of the works of Lazarus is another important framework which acknowledges the role
of resource loss and gains between the stressor-strain models. This theory also accepts that
stressor causes psychological resource depletion which affects strain. Moreover, this theory
posits that in order to prevent the resource loss caused by the stressor individuals’ builds
resource prevention strategies such as emotional self regulation in order to lessen the effects of
stressor on strain (Hobfoll, 1998; 2001; Schwarzer, 2001). Thus, both these theories highlight the
important role of both cognitions and emotions as parallel mechanisms through which stressor
causes strain in individuals.
7
The present study extends the literature on workplace bullying, stress and emotions by utilizing
the conservation of resources theory and appraisal theory of stress and coping as the overarching
theories to justify and interpret how and why workplace bullying as a stressor leads to negative
consequences i.e. strain. The current thesis attempts to bridge a significant gap in the workplace
bullying area by proposing two important mechanisms i.e. Perceived Support (Perceived
Organizational Support, Co-worker support and Supervisor Support) and Emotional dissonance
as mediators between workplace bullying and outcomes relationship. Moreover, the present
study aims to highlight an important individual difference variable i.e. Core Self evaluations as
an important moderator in the relationship between workplace bullying and emotional
dissonance and between workplace bullying and Perceived Support Types (POS, PSS and PCS).
1.2 Justification of the Study (Gap Analysis)
Although research examining workplace bullying has existed for the past many decades, before
the 1980’s bullying research was more subjective and descriptive in nature however, the ground-
breaking research initiated by Leymann (1990a) organized workplace bullying as a scientific and
methodically studied phenomenon (Zapf & Einarsen, 2005). From that time, interest and
attention in the area of bullying at the workplace has grown significantly.
Quantative research conducted in a number of countries such as USA, UK, Spain, France,
Germany, Canada, Nordic countries, Australia, India, Argentina, Hong Kong and Singapore etc
have provided support for the presence of bullying (Einarsen & Skogstad, 1996; Keashly &
Jagatic, 2000; Power, Brotheridge, Blenkinsopp, et.al., 2013), harmful effects on victims of
bullying (Hoel, Faragher, & Cooper, 2004; Leymann & Gustafsson, 1996; Mikkelsen &
Einarsen, 2002) work related (Salin, 2003a; Tepper, 2000; Vartia & Hyyti, 2002), and
8
personality variables (Aquino & Bradfield, 2000; Zapf, 1999) which may promote workplace
bullying. Although several studies have been conducted in the past on workplace bullying, there
are a lot of gaps which need to be investigated and filled to further clarify our understanding on
the complex phenomenon of workplace bullying.
Research studies have highlighted the significance of studying mediational processes in research.
For instance, it has been corroborated by researchers in the past that the findings of studies which
do not take into consideration the basic mechanisms causing such outcomes may be temporary
and short-lived (Mishra & Bhatnagar, 2010; Matheieu, DeShon, & Bergh, 2008). Nielsen &
Einarsen, (2012) in their recent meta-analysis on workplace bullying highlight that there is a
dearth of research in the literature on workplace bullying which has identified and explained the
mediating mechanisms as to how workplace bullying shows its harmful effects on outcomes.
Moreover, these authors argue that bullying might have secondary and indirect consequences on
employee's work related outcomes through the potential intervening mechanisms of various
variables which might further enhance this body of knowledge and increase the potency of
research on workplace bullying area. This study explores the mediational effects of emotional
dissonance and perceptions of support types on the relationship between workplace bullying and
job strains.
A recent meta-analysis has highlighted that work related bullying is inherently a psychological
experience and depends on the target’s subjective evaluation of being a victim or not (Vie,
Glaso, & Einarsen, 2011), therefore there is a need by future researchers to explore the
theoretical foundations of how victims’s experiences of bullying on well-being outcomes might
be explained through the cognitive mechanisms (Cassidy, McLaughlin & McDowell, 2014;
Nielsen & Einarsen, 2012; Samnani & Singh, 2012; Tuckey & Neall, 2014).
9
The present study employs two of the useful theoretical frameworks in the areas of stress namely
the appraisal theory of stress and coping and the conservation of resource theory as useful
underpinnings of how and through what mechanisms a stressor such as workplace bullying might
translate into negative consequences for the victim (Lazarus & Folkman, 1984; Lazarus, 1991;
Lazarus, 1999; Lazarus & Charash, 2001). Thus, the present study aims to extend these two
theoretical frameworks by examining the mediating mechanisms of emotional dissonance and
perceived support types between workplace bullying and job strains relationship.
This thesis also aims to extend the stressor-strain area by examining the role of cognitions and
emotions as parallel as well as the sequential mechanisms between workplace bullying and
strains. The literature on cognitions and emotions highlight many influential paradigms however
three dominant schools of thought stand out regarding the independence or interdependence of
cognitions and emotions. The first school of thought argues that one’s cognitions precede
emotions (Charland, 1997; Izard, 1992; Lazarus, 1999; Lazarus and Folkman, 1980, 1982).
The other school of thought considers emotions and cognitions as independent of one another i.e.
the rationalists view cognitions as part of the conscious mind where logic, information and
interpretation are formed of events vs. evolutionary psychologists view emotions as part of
subconscious mind. The evolutionary psychologist argues that emotions are involuntary and
reflexive particularly when individuals are faced with stressful events. These theorists give the
terminology of “Affective primacy” which refers to that emotional reactions are separate from
logical or cognitive assessments we do of other things, ideas or people (Tomkins, 1981, 1982;
Zajonc, 1980).
Lastly, the third school of path which is an extension of the above school of thought has given
the terminology of “twin path-way” model of emotions, cognitive-experiential model who
10
incorporate both the views i.e independence of cognitions and emotions as well as the causal
effect of cognitions on emotions that is cognitive evaluation of stressful events can determine
emotions and simultaneously when individuals are exposed to stimuli they might experience
affective states independent of cognitions (Lazarus, 1999; Ohman, 1999; LeDoux, 1996).
Therefore, this study builds on the previous research on stressor-strain and considers the parallel
as well as the causal role of individual’s cognitions i.e in the form of perceptions of support types
and emotions i.e. emotional dissonance between workplace bullying and job strains.
This study bridges the literatures on workplace bullying and emotional dissonance area by
highlighting workplace bullying as an important antecedent of emotional dissonance. Employees
in all kinds of organizations are subjected to the threat of negative behaviors at the workplace.
Past research on emotional labor or dissonance has mostly studied the outcomes of emotionally
dissonant employees (Ashforth & Humphrey, 1993; Grandey, 2000; Glomb & Tews, 2004).
However, a few studies have addressed what individual and organizational factors promote
emotional dissonance in employees (Mishra & Bhatnagar, 2010; Ma, Chen &Tang, 2010). One
recent study has suggested that employees who experience abusive supervision have higher
chances of performing surface acting which further creates work family conflict (Carlson,
Ferguson, Hunter &Whitten, 2012). So this study extends the research on emotional dissonance
by suggesting workplace bullying as a major factor generating emotional dissonance in
employees.
This study is unique in the sense that it contributes in the areas of workplace bullying and
emotional dissonance by applying the concept of emotional dissonance beyond service context to
organizational context and suggests that emotional dissonance emerges from workplace bullying
11
in employees. The research on emotional dissonance and emotional labor have been researched
with how employees indulge in emotional regulation while dealing with external customers (Ma,
Chen & Tang, 2010); it is argued by researchers that this line of investigation should also be
extended while interacting with other organizational members such as bosses, co-workers and
subordinates as well (Ashforth & Humphrey, 1993). Thus, the current research adds to the body
of knowledge on Emotional dissonance by proposing workplace bullying as an important factor
predicting employee’s emotional dissonance within organizational context.
The present thesis intends to extend the perceived organizational support literature by proposing
workplace bullying as an important stressor negatively affecting perceived organizational
support. Previous studies have suggested that bullying has the capacity to influence the
organizational climate rather being influenced by it (Skogstad, Torsheim & Einarsen, 2011).
Giorgi & Majer (2009) highlighted that bullying at the workplace is a kind of work related
endemic that might create a negative perceptions of organizational climate. Recently, Giorgi
(2012) found support for the argument that rather than organizational climate promoting
conducive environments for bullying to flourish, workplace bullying creates a negative and
stressful organizational climate for its employees.
A recent study on abusive supervision has established that workers who receive abusive
treatment from the supervisors are more likely to hold the organization responsible for such
behaviors and develop negative perceptions of organization support (Shoss, Eisenberger,
Restubog & Zagenczyk, 2013). The authors in a study of 3 samples reported that abusive
supervision was significantly negatively related to perceptions of organizational support which
further predicted employees’ retaliatory behaviors directed towards the organization, reduced job
12
performance and organizational citizenship behaviors (Shoss, Eisenberger, Restubog &
Zagenczyk, 2013).
Cassidy, McLaughlin & McDowell (2014) recently in an investigation of 2068 individuals
working in organizations in UK found support for the mediating roles of psychological capital
and social support as parallel mediators between experiences of workplace bullying and ill-
health. The authors contended that victims of bullying when exposed to maltreatment have
insufficient resources thus decreasing psychological capital and perceived social support (family,
friends and colleagues) which ultimately depletes their health and well-being. Extending this
limited line of inquiry on workplace bullying, this study answers the recent calls by researchers
to link workplace and perceived support and aims to consider workplace bullying as an important
predictor of perceived support types (POS, PCS and PSS).
This study adds important insights into the role of dispositional variable in the workplace
bullying process. It has been contended that an individual’s personality traits may contribute as a
significant factor in affecting a person's appraisal of the potential stressors as well as the
individual’s perceived coping capacity which either increases or decreases stress reactions
(Cohen & Edwards, 1989; Cox & Ferguson, 1991; Spector, Zapf, Chen & Frese, 2000).
Moreover, it has been asserted by researchers in the past that future researchers should examine
personality characteristics that moderate the association between bullying and the various kinds
of individual processes and reactions (Vartia, 2001). The present study contributes to the
literature examining the dispositional variables in the workplace bullying area by taking core self
evaluations as an important moderator between workplace bulling and perceived support types
and between workplace bullying and emotional dissonance.
13
1.3 Significance of the Research (Theoretical Contribution)
The present study contributes to the literatures on workplace bullying, emotional dissonance and
perceived support by employing two well known theoretical frameworks which include
Cognitive appraisal theory of stress and coping and Conservation of Resource theory as broader
theories to justify the proposed links. According to the appraisal theory of stress and coping,
stress occurs depending on how a stressor is evaluated by an individual and how an individual
assess his or her resources to deal with the stressor. Therefore, individual’s cognitive appraisals
and subjective evaluations of the stressor and their emotional coping reactions play an important
role in determining the outcomes of stressful interactions at the workplace (Vie, Glaso, &
Einarsen, 2011).
The present investigation adds to the cognitive appraisal theory of stress and coping by
proposing two important mediators in the workplace bullying and outcomes relationship. In line
with this theory, it is the contention of the present study that workplace bullying leads to strains
through the mechanisms of perceived support types. Moreover, this study extends this theory by
proposing an important mediator of emotional dissonance between workplace bullying and job
strains (Job stress, Job burnout, Psychological strain and Intensions to quit).
This study adds to the literatures on workplace bullying, perceived support and emotional
dissonance in the perspective of the conservation of resource theory. This theory asserts that
stress emerges because of three conditions i.e. when individual’s important resources are
susceptible to resource loss, when resources are actually lost because of stressful circumstances
and when individuals are unable to increase resources after major resource investment. Job
burnout and stress are two outcomes of the stressful circumstances that normally evolve from a
14
slow depletion of resources without increasing resources or a replacement of resource loss. By
employing this theoretical framework, this study sheds light on how a stressor such as workplace
bullying creates resource loss and translates into strain through the mechanisms of lower
perceptions of support and emotional dissonance.
Einarsen et. al. (2003) advanced an explanatory model to conceptualize the phenomenon of
workplace bullying which intends to offer assistance for further research and interventions in this
area. Their theoretical framework proposes individual and organizational factors as predictors of
bullying, personality traits and situational factors as moderators and individual's
cognitions/interpretations and responses as potential mediators and employee/organizational
consequences as outcomes in the bullying model. Although the predictors and consequences of
workplace bullying have been widely researched, however future research is vital to scrutinize
why and through what processes bullying creates harmful consequences on an individual’s
psychological and emotional well-being (Sandvik, Tracy & Alberts, 2007; Nielsen & Einarsen,
2012). Thus, this study attempts to highlight and explain how and why workplace bullying
manifests itself into higher stress, burnout, psychological strain and intensions to quit.
It has been argued by researchers that due to a lack of longitudinal research designs and
inconsistent relationships between workplace bullying and outcomes, most models explaining
bullying have given support to organizational culture and employee’s perceptions of
organizational climate as antecedents promoting bullying at the workplace rather than a
consequence of bullying (Kivimaki, Vahtera, Pentti & Ferri, 2000; Pfeffer, 2007; Vardi & Weitz,
2004). However, recent studies have suggested that it is highly appropriate that instances of
reverse causality and interdependency case might exist which means that instead of
organizational climate promoting bullying; casual effects of bullying promoting negative
15
perceptions about organizational climate and support which needs to be investigated by future
researchers (Giorgi and Majer 2009; Giorgi, 2012). The present research study adds to the
workplace bullying literature by viewing workplace bullying as an important stressor negatively
influencing perceived organizational support.
The present investigation attempts to link the workplace bullying and emotional dissonance
literatures by introducing workplace bullying as a key determinant of emotional dissonance and
by applying the concept of emotional dissonance beyond service context to organization context.
It has been asserted by researchers in the emotional labor area that future investigations should
highlight the degree to which the concept of emotional dissonance can be applied beyond service
roles to other organizational roles such as while dealing with supervisors, colleagues and
subordinates (Ashforth & Humphrey, 1993; Rafaeli & Sutton, 1989). Such organizational roles
also require a set of social and organizational expectations to be fulfilled by employees and
emotions are also experienced when executing such organizational roles it is difficult to
conceptualize an organizational role to which display rules would not apply. Thus, the present
study extends the emotional dissonance area by identifying workplace bullying as an important
predictor of emotional dissonance thereby applying it in a variety of settings and in interaction
with a variety of organizational roles.
1.4 Research Problem
Although the past literature in the area of workplace bullying shows that workplace bullying is a
severe stressor which creates strain however the underlying mechanisms through which
workplace bullying shows its deleterious effects on job strains are still unclear and unexplained
(Monks, Smith, Naylor, Barter, Ireland & Coyne, 2009). As mentioned, although quantitative
16
studies have been conducted in a number of countries to investigate the various antecedents,
moderators and consequences however, studies in the area of workplace bullying in the context
of Pakistan are scarce particularly there is a need to understand and explain the how's and why
through which workplace bullying manifests into job strains in employees (Cassidy, McLaughlin
& McDowell, 2014; Einarsen et. al., 2003; Nielsen & Einarsen, 2012; Samnani & Singh, 2012;
Tuckey & Neall, 2014).
Particularly, there is a need to answer the recent calls by researchers to explain how workplace
bullying as a stressor might affect perceived support and emotional dissonance which translates
into job strains (job stress, job burnout, psychological strain and intensions to quit). Moreover, as
explained previously there is a need to highlight personality characteristics in the relationship of
work related bullying and various kinds of individual processes and reactions (Vartia, 2001;
Nielsen & Einarsen, 2012; Samnani & Singh, 2012; Tuckey & Neall, 2014). Thus, this study
takes into account to what extent personality traits (i.e. Core self evaluation) moderates the
relationship between workplace bullying and emotional dissonance on one hand and workplace
bullying and perceived support types on the other.
1.5 Research Questions
The current study's research questions are:
1. How does workplace bullying influences employee’s outcomes (Job stress, Job burnout,
Psychological Strain and Intensions to Quit)?
2. How does workplace bullying impacts emotional dissonance in employees?
17
3. What is the effect of workplace bullying on employee’s Perceived support types
(Perceived Organization support, Perceived Supervisor support and Perceived Coworker
support)?
4. What is the impact of emotional dissonance on employee’s outcomes (Job stress, Job
burnout, Psychological Strain and Intensions to Quit)? What is the impact of perceived
support types on employee’s outcomes?
5. How does employee’s emotional dissonance and perceived support types mediates the
relationship between bullying and employees’ job strains?
6. How does core self evaluation acts as a moderator in the relationship between workplace
bullying and Emotional dissonance and between workplace bullying and perceived
support types?
1.6 Research Objectives
The current study's research objectives are:
1. To investigate the impact of bullying on employee’s job strains (Job stress, Job burnout,
Psychological Strain and Intensions to Quit).
2. To examine the impact of workplace bullying in creating emotional dissonance in
employees.
3. To investigate the influence of workplace bullying on employee’s perceived support
types (Organization support, Supervisor support and Coworker support).
18
4. To analyze the influence of Emotional dissonance in creating employee’s strains (Job
stress, Job burnout, Psychological Strain and Intensions to Quit). To examine the
relationship between perceived support types on employee’s outcomes.
5. To explore the mediating effects of emotional dissonance between the workplace
bullying and job strains relationship and investigate the mediating role of support types
between the workplace bullying and job strains relationship.
6. To study core self evaluation (CSE) trait as a moderator in the relationship between
workplace bullying and emotional dissonance. To analyze the moderating role of CSE as
a personality trait between workplace bullying and perceived support types.
1.7 Research Model
The following figure below shows the research model which is deduced on the basis of the
above research gap.
19
Independent Variable Mediating Variables Dependent Variables
Moderating Variable
H7
H3
H10
H5
H9 (a, b,c)
H8 H1
H6 (a, b, c)
H2(a, b,c)
H4 (a, b, c)
Figure 1. Research Model
Figure depicting (a) the direct effects of workplace bullying on job strains (Job stress, Job burnout, Psychological
strain and intensions to quit); (b) the (simple, parallel and sequential) mediating effects of Emotional dissonance and
Perceived support types between workplace bullying and Job strains and (c) CSE personality trait as moderator
between workplace bullying and emotional dissonance on one hand and workplace bullying and perceived support
types on the other.
Emotional
Dissonance
Perceived Support
Perceived Supervisor
Support
Perceived Co-worker
support
Perceived Organization
Support
Job Stress
Core Self Evaluations
Workplace Bullying
Job Burnout
Psychological Strain
Intensions to Quit
20
CHAPTER 2
LITERATURE REVIEW
2.1 Workplace Bullying
Bullying at the workplace has been examined and studied under a variety of labels. For example,
in UK, the label bullying is commonly used. Some researchers believe that the term bullying
invoke interpretations of physical hostility which is uncommonly indicated in bullying at the
workplace (Leymann, 1996). In a few European countries and Germany (Zapf & Einarsen, 2011)
the term mobbing is used to describe such aggressive behaviors whereas in the US similar
concepts have been labeled as abusive supervision (Tepper, 2000), Harassment at the workplace
(Brodsky, 1976) and emotional abuse (Keashly & Harvey, 2005). Even though many different
terminologies have been employed to define the concept of workplace bullying but normally
there is agreement among researchers on a common operationalization of bullying.
Workplace bullying is defined as an escalating process which occurs when an individual is
repeatedly and continuously exposed to negative acts from one or more individuals, which range
over duration (i.e. about six months), in which there is a power imbalance and where the victim
due to a numbers of reasons feels helpless against such behaviors and is unable to protect
himself/herself (Einarsen, 2000; Einarsen et al., 2011; Salin, 2003; Zapf, Einarsen, Hoel, Vartia,
2003).
21
Einarsen and Hoel (2001) conducted exploratory and confirmatory factor analysis of the NAQ
(Negative Acts Questionnaire) which resulted in a two factor solution comprising of individually
directed bullying and work oriented bullying. Bullying behaviors directed towards one's work
includes actions such as setting unrealistic deadlines, impossible workloads, stringent checks and
balances and hiding pertinent information from the victim. Whereas personal bullying
encompasses negative behaviors intended to create more personal attacks such as gossiping,
humiliating comments, making fun of and continuously criticizing the victim (Einarsen & Hoel,
2001; Einarsen & Mikkelsen, 2000; Neuman & Baron, 1997; Nielsen et al. 2009).
2.2 Emotional Dissonance
Hochschild (1983) in his pioneer work introduced the concept of Emotional labor to explain how
individuals manage their feelings at the workplace so that they can attain their performance
goals. The concept of emotional labor has been usually been prevalent and researched in the
service sector where individual communications have a significant contribution in their work
(Ma, Chen &Tang, 2010). It is the contention of the present study that emotional labor or
dissonance is pertinent for all kinds of jobs and is required for all those jobs which require
interaction with significant others at the workplace.
Emotional labor is defined as the purposeful preparation, effort and management of work related
feelings and emotions during interpersonal interactions (Morris & Feldman, 1996). This concept
of emotional labor has been explained through the emotional regulation model which explains
how employees use emotional regulation to achieve emotional labor (Grandey, 2000). Two
processes of emotional labor have been contended in past research i.e. surface acting or
emotional dissonance and Deep acting (Hochschild, 1979; 1983). Surface acting means that
22
employees display fake emotions which are required by the organization even though they are
not genuinely feeling them. For example, employees might demonstrate a fake smile even when
they are in a bad mood or when they are dealing with problematic customers. Whereas deep
acting is the process of modifying or changing one’s inner thoughts to match with the display
rules (Brotheridge, 2002).
Emotional dissonance is a stress reaction which encompasses a discrepancy between employee’s
inner emotions and expressed emotions in accordance with organizational rules and have been
described as a person role conflict because the individual does not associate with one’s role and
job requirements and must change their response in order to meet and satisfy current job and role
expectations (Rafaeli & Sutton, 1987; Abraham, 1999).
Emotional dissonance has been conceived in a number of different ways in the literature on
emotional labor. Some authors consider emotional dissonance as a consequent state which
emerges as a result of emotional labor i.e. when employees display emotions as a part of their job
requirement which do not match with their inner felt emotions (Diefendorff et al., 2005; Heuven
& Bakker, 2003). Others have viewed emotional dissonance as a cause of emotional labor where
the inconsistency between the expressed and experienced emotions affects emotional labor
(Grebner, Semmer, Faso, Gut, Kalin & Elfering, 2003; Zapf & Holz, 2006). Lastly, some studies
have considered emotional dissonance as one of the types of emotional labor (Morris &
Feldman, 1996).For the current thesis, emotional dissonance is operationalized as a condition of
conflict, tension and discomfort which happens due to the inconsistency between inner and
exhibited emotions or between genuinely experienced and portrayed emotions (Mishra &
Bhatnagar, 2010).
23
2.3 Perceived Support
2.3.1 Perceived Managerial/Supervisor Support
Perceptions of Supervisor support is defined as the extent to which individuals form overall
perceptions regarding the extent to which the managers/bosses recognize their efforts and are
concerned about their well-being (Kottke & Sharafinski, 1988).
2.3.2 Perceived Co-worker Support
Perceived coworker support refers to the extent to which individual’s view that their coworkers
would be beneficial in providing help and support in case of any need such as job related matters
and will play an active role in providing emotional support to them as well (Thoits, 1985).
2.3.3 Perceptions of Organizational Support
Eisenberger, Huntington, Hutchison & Sowa (1986) initiated the idea of perceptions of
organizational support as a means of understanding employee’s evaluations regarding the
organization's dedication towards them. Perceived organizational support construct encompasses
individual’s overall assessments regarding the extent to which the organization supports them, is
concerned about their welfare, value their individual as well as organizational efforts and is fair
and just in identifying and rewarding extra effort and higher dedication (Eisenberger et.al., 1986;
Eisenberger, Fasolo & Davis-LaMastro, 1990; Rhodes & Eisenberger, 2002; Shore & Shore,
1995).
24
2.4 Personality
2.4.1 Core Self Evaluations
Core Self evaluations refers to an individual’s underlying assumptions about himself or herself.
These assumptions are dependent on one’s basic values, beliefs and principles which influence
the overall level of well-being and self- worth. Judge et al. (1998) in his pioneer work introduced
the variable of CSE which consists of one’s Self-efficacy, Self-esteem, Neuroticism and Locus
of control.
Self esteem is the degree to which persons like themselves and consider themselves as
competent, important, successful and worthy (Coopersmith, 1967). Generalized self efficacy is a
person's trust in one's abilities of showing higher performance across a wide variety of situations.
Locus of control is a person's confidence in one’s capacity to control one’s external factors
(Judge, Locke & Durham, 1997). Neuroticism is an individual’s deficiency in emotional
consistency and one's propensity to view events and situations in a negative light (Judge, Vianen
& De Pater, 2004).
Judge, Erez, Bono & Thoresen (2002) asserted that one should conceptualize CSE in terms of a
broader construct consisting of shared variance of its various dimensions. Moreover, past results
of confirmatory factor analysis has clearly resulted and supported for a one factor structure of
CSE (Judge et.al., 1998; 2000).
Successive research analyzing the construct validity of CSE favors this concept as all the four
dimensions are interrelated and demonstrate similar relationships with outcomes for instance job
performance, job and life satisfaction (Bono & Judge, 2003; Judge et.al., 2002).
25
2.5 Job Strains
2.5.1 Job Stress
Job Stress refers to an unfavorable physical and emotional state which occurs as a consequence
of negative incidences at the workplace, specifically events that are uncertain or complex to
handle by the individual (Beehr & Bhagat, 1985; Hart & Cooper, 2001). The work related
incidents which create stress are called stressors whereas the outcomes of stress either in terms of
individual's health or behaviors are called strains (Hart & Cooper, 2001).
2.5.2 Job Burnout
Burnout refers to one's reaction to persistent interpersonal and emotional stressors at the work
which is classified as cynicism, emotional exhaustion and diminished feelings of efficacy at
work (Maslach, Schaufeli & Leiter, 2001). Emotional exhaustion is defined as when one
develops feelings of being constantly used up or fatigued as a result of work and devoid of
mental, physical and emotional resources (Maslach, Schaufeli & Leiter, 2001). Cynicism refers
to a state whereby an employee distance itself from the workplace which becomes evident in the
form of an immediate, negative and indifferent attitude towards different feature’s of one’s job
qualifies the emergence (Maslach et.al., 2001). Reduced feelings of efficacy refers to involve an
individual’s assessment of one’s past, present and future accomplishments on job which
encompass feelings of an inability to perform one’s work related demands (Maslach et al., 2001).
2.5.3 Psychological Strain
Psychological strain refers to the extent to which an individual faces psychological and
emotional health complaints (Osipow, 1998). Psychological strain depicts the long term negative
26
effects on the employee due to constant experience of work related stressors and consists of such
indicators as anxiety (Billings & Moos, 1982), sadness (Tepper, 2001; Zohar, 1995) and social
dysfunctions (Banks et al., 1980).
2.5.4 Intensions to Quit
Behavioral and voluntary withdrawal behaviors are often subsumed under the heading of
turnover intensions or intensions to quit. Turnover intensions is defined as the degree to which an
employee wants to stay or leave with the organization (Cotton & Tuttle, 1986).
2.6 Theoretical Background
2.6.1 Conservation of Resource Theory
Hobfoll’s (1988, 1989, 1998a) conservation of resources theory asserts that individuals struggle
to acquire, preserve, defend, and increase valued resources and minimize any threats of resource
loss. Resources are defined as an employee’s overall capacity to meet his/her essential needs
(Hobfoll, 2002). Resources might include one’s physical resources such as bodily power,
monetary assets (earnings and possessions), societal resources (support, connections), mental
possessions (abilities, skills & Knowledge) and motivation related resources i.e. goal
involvement and self efficacy (Wang, 2007).
This conservation process consists of two equivalent mechanisms. The accumulation mechanism
is defined as a means by which employees utilize their resources to regulate their behaviors and
gain control over the environment to build up their resources to meet their needs (Hobfoll, 2002).
The protection mechanism emphasizes individual’s capacity to defend, guard and prevent
resource losses which includes primary and secondary resource loss.
27
Hobfoll (1988, 1989, 1998a) in his theory emphasizes the assumption of primacy of loss which
means that resource losses are more important than resource gains. When individuals encounter
stressful or demanding situations it may lead to loss of resources which is called primary
resource loss. Following primary resource loss, resource investment is required on the part of
individuals which mean that individuals must devote a considerable amount of resources so as to
defend against resource depletion, recoup from and increase one’s resources. If additional
resources invested tends to further exhausts individual’s resources secondary resource loss
occurs. As a result of the secondary resource loss individuals become inefficient and utilize
ineffective loss control strategies which then leads to a loss spiral in which resources needed to
regulate emotions and behavior become drained and used up even more rapidly (Bacharach &
Bamberger, 2007; Hobfoll, 2002). These two principles of primary and secondary resource loss
emphasize that resource gain and loss spirals respectively emerge in stressful or demanding
situations at the workplace or where individuals or organizations are resource deficient (Hobfoll,
2010). In other words, individuals with higher resources are less susceptible to resource
exhaustion and have increased capacity to increase resources. On the other hand, individuals who
possess limited resources are highly susceptible to resource loss and are not as much able to
increase one’s resources in the face of unfavorable or demanding situations as compared to their
counterparts who have higher resources.
2.6.2 Cognitive Appraisal theory of Stress and Coping
Lazarus and Folkman (1980) in their pioneer work introduced the cognitive appraisal theory of
stress and coping into the stress and emotions literature and more recently this theory has been
expanded as the cognitive motivational relational theory (Lazarus, 1991, 1999). This theory
28
focuses on the constant and joint nature of the interaction between the individual and the
environment. This model since its inception has not only been further developed and advanced,
but it has also been extended to a broader theoretical framework of emotional and coping
processes (Lazarus, 1966; Lazarus, 1991).
The appraisal theory of stress and coping asserts that stress emerge from an individual’s
subjective evaluations of the encountered situations and their influence on important goals
(Frijda, 1986; Lazarus, 1991; Roseman, Spindel & Jose, 1990). Smith & Lazarus (1993)
suggested two classification of appraisals that affect emotions i.e. primary and secondary
appraisal. In primary appraisal individual evaluate how much significance they attach to an event
(motivation significance) and the degree to which this situation is fruitful for maintaining one’s
individual goals (motivational congruence). According to Smith and Lazarus (1993) unfavorable
situations hold a high degree of motivational relevance but a low level of motivation congruence.
Secondary appraisal assess an individual’s survival or coping processes available and includes
problem focused vs. emotion focused coping. Problem focused coping deals with generating
alternatives which will influence the situation and emotional focused coping consists of one’s
capacity to affectively alter according to the circumstances. Secondary appraisal in combination
with primary appraisal determines the emotions generated (Griner & Smith, 2000).
Lazarus (1991) conceptualizes this model as a broader framework and a complex mechanism of
emotions and stress consisting of various predictors, mediating processes and outcomes.
Antecedents are individual’s resources such as money, social connections, abilities, beliefs on
one hand and quantitative burden, decisive incidents, stressful situations or contextual limitations
on the other hand. Mediating processes imply cognitive evaluation of such resources and
demands as well as the coping attempts to overcome or at least survive in such situations. The
29
outcomes or negative consequences of appraisal on one hand and coping on the other hand create
long term effects for individuals such as impaired physiological, psychological stress and
burnout and social functioning.
2.7 Theoretical Framework and Hypotheses Development
2.7.1 Workplace Bullying and Job Strains relationship (Job Stress, Job Burnout,
Psychological Strain and Intensions to quit)
It is commonly acknowledged by researchers in the workplace bullying area that bullying is an
important workplace stressor. Past studies of workplace bullying have examined the harmful
effects of bullying on employee’s physiological, cognitive and emotional outcomes (Zapf et.al.,
2003). A considerable amount of research has found that people who face bullying at the
workplace suffer from negative well-being (Hoel et.al., 2004; Niedl, 1996; Tepper, 2000). For
instance numerous studies have found that the victims of workplace bullying suffer from high
degree of stress, health illness and burnout (Leymann & Gustafsson, 1996; Matthiesen &
Einarsen, 2004; Tehrani, 2004).
Moreover, it has also been substantiated by researchers that even the non-bullied employees who
observe other individual’s being bullied report high levels of general stress, burnout and
psychological reactions than do employees who exist in organizations without bullying (Vartia,
2001). Therefore, workplace bullying is not only a serious concern for the victims of bullying but
also a pervasive issue which severely impairs the health and welfare of all the individuals
working in the organization. Thus, it is hypothesized:
30
Hypothesis 1: Workplace Bullying is positively related to Job Stress, Job Burnout, Psychological
Strain and Intensions to quit.
2.7.2 Relationship of Perceived Support types and Job Strains (Job Stress, Job Burnout,
Psychological Strain and Intensions to quit)
The literature on organization support clearly depict that perceived supervisor support is related
to a number of outcomes both individual and work related. For instance, it has been corroborated
in a number of studies that workers who view lower supervisor support have higher chances to
experience psychological strain and job stress (Bratt et al. 2000; Carlson and Perrewe, 1999;
Dormann & Zapf, 1999).
In addition, Perceived Coworker support has received considerable attention from researchers for
the past many decades and is significantly associated with employee’s well-being such as lower
stress, burnout and psychological strain (Fisher, 1985; Van Emmerik, Euwema & Bakker, 2007).
Past studies examining the effects of perceptions of organizational support on employee's
outcomes suggest that employees who develop lower perceptions of organizational support have
greater incidences to develop strain outcomes such as fatigue (Cropanzano, Howes, Grandey &
Toth, 1997), job burnout (Cropanzano et.al., 1997), nervousness and depression and higher
intensions to quit (Robblee, 1998; Venkatachalam, 1995). It has also been substantiated in past
research that negative perceptions of organizational support increase one’s stress level as well
(Viswesvaran, Sanchez, & Fisher, 1999). Thus, it is hypothesized:
Hypothesis 2a: Perceived Organization Support is negatively related to Job Strains (Job Stress,
Job Burnout, Psychological Strain and Intensions to quit.
31
Hypothesis 2b: Perceived Coworker Support is negatively related to Job Strains (Job Stress, Job
Burnout, Psychological Strain and Intensions to quit.
Hypothesis 2c: Perceived Supervisor Support is negatively related to Job Strains (Job Stress, Job
Burnout, Psychological Strain and Intensions to quit.
2.7.3 Relationship of Emotional Dissonance and Job Strains (Job Stress, Job Burnout,
Psychological Strain and Intensions to quit)
Past research suggests that organizations might benefit from employees’ emotional dissonance
however emotional dissonance has negative consequences for employees’ well-being
(Hochschild, 1983; Grandey, 2000; Zapf et.al., 1999). Research has identified that employees
who experience emotional dissonance are more likely to encounter negative outcomes such stress
and burnout (Brotheridge & Grandey, 2002; Hochschild, 1983; Morris & Feldman, 1997). A
meta-analysis examining the consequences of emotional labor have identified that employees’
emotional dissonance is associated with a number of job related and well-being consequences
such as strain and higher intensions to quit (Mishra & Bhatnagar, 2010;
Particularly, Job burnout is one of the frequently experienced and examined outcomes of
emotional dissonance in past research that has been tested and supported in a broad range of
service oriented jobs (Abraham, 1998; Brotheridge & Lee, 1998; Heuven & Bakker, 2003; Zapf
et.al., 1999). Moreover, emotional dissonance has also been found to be strongly associated with
all three types of job burnout i.e. emotionally dissonant employees because of incompatible
displayed and true emotions might suffer from high levels of cynicism (Brotheridge & Lee,
2002; 2003; Naring, Briet & Brouwers, 2006), emotional exhaustion (Brotheridge & Lee, 2003;
32
Golderberg & Grandey, 2007; Martinez-Inigo et.al., 2007; Naring et.al., 2006; Totterdell &
Holman, 2003) and diminished self efficacy (Brotheridge & Lee, 2003). Thus, it is hypothesized:
Hypothesis 3: Emotional Dissonance is positively related to Job Stress, Job Burnout,
Psychological Strain and Intensions to quit.
2.7.4 Relationship of Workplace Bullying with Perceived support types (POS, PSS & PCS)
Although majority of studies have examined how negative work climates might foster
circumstance for workplace bullying to emerge (Agervold, 2009; Agervold & Mikkelsen, 2004;
Hoel & Cooper, 2000; Vartia, 1996; Zapf et.al., 1996) only one recent study has implicitly
addressed how being subjected to bullying might lead to the development of negative
organizational climate (Giorgi, 2012). Therefore, research focusing on the association between
work environmental factors and bullying have not analyzed the critical concern in today’s
workplace of whether perceptions of work environment might generate conducive environments
for bullying to take place or whether employees who are bullied are more likely to develop
negative perceptions about the organizational climate particularly about perceptions of
organizational support (Skogstad, Torsheim & Einarsen, 2011).
According to the recent literature, the concept of workplace bullying as an antecedent to
organizational culture and climate perceptions seems more rational (Giorgi & Majer 2009;
Hauge et al., 2011). Secondly, more theoretical and empirical research is needed that can
substantiate the impact of workplace bullying on employee’s perceptions of organizational
climate and culture (Giorgi, 2012; Hauge et al., 2011). The present study takes up the recent calls
made by researchers in this regard and argues that workplace bullying is an important predictor
33
of perceived organizational support. By employing evidences from past research and by
employing the two theoretical pillars of this study namely cognitive appraisal theory of stress and
conservation of resource it is suggested that workplace bullying leads to the development of
lower perceptions of organizational support.
Research examining the antecedents of perceived organizational support highlight that different
forms of justice particularly procedural and interactional justice have the capacity to increase
one’s perception’s of organizational support. This line of inquiry argues that when employees are
given information regarding how their rewards are established and when they are dealt with
politeness and esteem they are greater chances that these individuals develop higher perceptions
of organizational support (Rhoades & Eisenberger, 2002).
Although past research has established a positive association between fair treatment and
perceived organizational support (Rhoades & Eisenberger, 2002) less is known about how
unfavorable and negative treatment received by relevant others has an impact on perceptions of
organization support (Shoss, Eisenberger, Restubog & Zagenczyk, 2013).
Recent research has highlighted that victims of abusive supervision hold the organization
responsible for the actions of the supervisor and bosses (Shoss, Eisenberger, Restubog &
Zagenczyk, 2013). Levinson (1965) asserted that since the organization is ethically and officially
accountable for the behavior of its managers/coworkers in their position of guiding and assessing
the subordinates, therefore employees mostly place the blame for abusive treatment to the
organization. Moreover, Cassidy, McLaughlin & McDowell (2014) recent research also suggests
that employees who experience a stressful and demanding situation such as bullying also are
34
more likely to perceive less social support (i.e. from family, friends and co-workers) which
might translate into poorer health and well-being outcomes for its victims.
From the perspective of employees, managers/bosses as well as coworkers are operating in the
organization environment which has empowered them with various positions to fulfill their
responsibilities. If these individuals operating under the boundaries of the organization engage in
harassing and disgracing behaviors towards the employees, the victims would feel that the
organization has intentionally and deliberately allowed these individuals to continue such
behaviors resulting in the development of lower perceptions of organization support. Moreover,
this might also suggest to the targets of such abusive behaviors that the organization is unable to
take remedial measures to reduce or stop the abuse from continuing thus creating lower
perceptions of organization support.
Moreover, since managers/bosses work as organizational representatives therefore when
employees receive positive treatment from a boss they have increased chances to develop higher
interpretations regarding organizational support towards them of organizational support whereas
unfair and humiliating treatment might signal to them that they are not valued by the
organization (Eisenberger et al., 1986; Eisenberger, Stinglhamber, Vandenberghe, Sucharski &
Rhoades, in press). It is contended that when employees are bullied and subjected to negative
treatment they feel that organization is least bothered about their welfare and these negative
events are consciously or intentionally performed to harm them leading to lower perceptions of
organizational support. In addition, these victims subjected to emotional abuse feel that co-
workers are not supporting and defending them whereas supervisors are turning a deaf ear
towards them thus generating lower perceptions of supervisor and co-worker support.
35
The workplace bullying and perceived organizational support relationship can also be explained
through the lens of cognitive appraisal theory of stress. Cognitive appraisal theory of stress and
coping highlights two critical factors i.e. an individual’s cognitive evaluations and coping
mechanisms as mediators between stressor-strain relationships (Folkman, Lazarus, Schetter,
DeLongis & Gruen, 1986).
According to this theory, cognitive evaluation is a mechanism through which the individual
assess whether a specific encounter with the environment is important to his or her well-being
and if yes then in what ways. In primary appraisal, an individual assess whether he or she has
anything to lose as a result of the stressful encounter. In the context of workplace bullying it is
asserted that primary appraisal occurs when an individual who experiences hostile and
aggressive treatment in the form of bullying start to cognitively evaluates the situation and
perceive lower organizational support since these individuals believe that bullying is harmful in
the achievement of goals and their self esteem is also at a threat.
This cognitive assessment of a demanding event such as bullying creates lower perceptions of
organizational support as when employees are intimidated and abused these employee feel that
the organization is not taking any action and organization does not value them as an integral part
of the organization.
The agent system model also sheds light on how workplace bullying might lead to the
development of lower perceptions of supervisor/coworker support. This model asserts that based
on the fair/unfair treatment received from the supervisor/coworkers individuals reciprocate and
direct their behaviors towards the source (Masterson, Lewis, Goldman, & Taylor, 2000). In this
context, I believe that when employees experience workplace bullying they might blame the
36
target of these negative actions and hold the supervisors/coworkers accountable for such
mistreatment thus creating negative perceptions of supervisor/coworker support.
The agent system model contends that interactional injustice is a potent predictor of employees
reciprocating actions (Folger & Cropanzano, 1998). In the perspective of workplace bullying this
model would assert that when employees face workplace bullying they hold the individuals
responsible for deliberately defying the standards for suitable behavior thus developing lower
perceptions of supervisor and coworker support.
Moreover, when employees are exposed to workplace bullying they might attribute the blame for
such an abusive treatment towards the supervisor/coworkers and feel that these individuals do
not care about their well-being thus promoting lower perceptions of supervisor/coworker
support. Therefore from the above empirical evidences and theoretical justification, it is
suggested that workplace bullying would create lower perceptions of support as individuals
would view the environment as destructive where the organization is not doing much to curtail
the harmful acts of bullying and simultaneously feel that the supervisors/coworkers are also not
supportive towards the employee. Thus, it is hypothesized:
Hypothesis 4a: Workplace Bullying is negatively related to Perceived Organization Support.
Hypothesis 4b: Workplace Bullying is negatively related to Perceived Coworker Support.
Hypothesis 4c: Workplace Bullying is negatively related to Perceived Supervisor Support.
37
2.7.5 Relationship of Workplace Bullying and Emotional Dissonance
The impact of workplace bullying on emotional dissonance is suggested by utilizing the
evidences made from past research and in the light of appraisal theory of stress and coping,
conservation of resources theory and theory of façade creation.
Emotions are a sign of one’s essential, important and valued identity. Emotions are often
involuntary and are influenced by work related stressors, mood, fatigue apart from expected
work demands (Thoits, 1990). It has been established in the research on emotional dissonance
that customer service employees in particular and employees in general are required to comply to
organizational display rules (Ekman & Friesen, 1969) which specify either directly or indirectly
that they should express positive and suppress negative emotions (Brotheridge & Grandey, 2002;
Hochschild, 1983; VanMaanen & Kunda, 1989). However, employees throughout their
workdays naturally experience a number of positive and negative emotions depending on their
exposure to organizational and work stressors. These employees are expected to comply with
organizational rules while displaying emotions of any kind and when performing their work
duties and responsibilities (Scott & Barnes, 2011). However, it might happen that emotions
which are internally felt are not the same as the ones which are displayed while interacting with
others depicting emotional dissonance.
Past research linking the workplace bullying and emotional dissonance concepts have analyzed it
through the lens of customer aggression creating emotional dissonance in employees (Grandey et
al., 2004). This line of research contends that employees who evaluate aggressive customers as
highly traumatic reported high levels of emotional dissonance than those who found them mildly
stressful (Bedi & Schat, 2007).
38
Applying this argument to workplace bullying, I contend that employees who encounter bullying
at the work might deliberately and consciously attempt to change their emotions creating
emotional dissonance. This faking and suppressing of emotions when dealing with abusive
others may make them feel false and insincere. It is the assertion of the present study that this
emotional dissonance or masking of one’s genuinely felt emotions is more likely to occur as a
result of stressful and demanding work environment where bullying is prevalent.
It has also been substantiated in the research on emotional labor that managing emotional
demands at work could be made possible and emotional dissonance is lessened by supportive
colleagues and social harmony at workplace (Luchman & Gonzalez-Morales, 2010; Van
Dierendonck, Schaufeli & Buunk, 1998). Extending this argument for workplace bullying, it is
believed that a workplace rife with bullying creates an environment where employees feel
emotionally abused and battered which results in employees suppressing their true emotions and
unable to genuinely feel what they display resulting in higher levels of emotional dissonance.
The COR theory (Hobfoll, 1989; 2002) also adds important insights into how workplace bullying
might lead to emotional dissonance in employees. As explained previously, this theory highlights
that individuals who have deficient resources are the most susceptible to resource loss. It is
believed that when employees face bullying and hostile behaviors at the workplace resource loss
occurs which leaves them with little emotional and psychological resources to deal with such
incidents. Bullying and aggressive treatment creates emotional dissonance in them as in order to
survive in the workplace they indulge in emotional regulation strategies by suppressing true
emotions and portraying acceptable emotions which makes them vulnerable to further resource
loss. Thus, bullying incidences promote resource loss in employees who then with poorly
39
available and deficient resources indulge in self coping strategies leading to high levels of
emotional dissonance in them.
The cognitive appraisal theory of stress and coping is another relevant theory which supports my
assertion of workplace bullying promoting high levels of emotional dissonance in employees.
According to this theory, primary appraisal occurs when individuals who face stressful situations
start developing perceptions and cognitions regarding how much harmful that event is for the
individual. Once individuals have cognitively established that certain situations are harmful for
them, secondary appraisal process commences. In secondary appraisal process, individuals
assess what if anything can be done to stop or avert back the damage done because of the
stressful event. Primary and secondary evaluation in combination determine whether the
encounter with the stressful event and individual’s subsequent cognitive evaluation and resulting
coping mechanisms are important for one’s well-being and if so then whether it is perceived to
be threatening i-e those situations which include the likelihood of creating harm or challenging i-
e situations which are likely to be viewed as an opportunity or as a benefit (Folkman, Lazarus,
Schetter, DeLongis & Gruen, 1986).
It is contended that secondary appraisal occurs when employees who face workplace bullying
think about the various coping options available to deal with such a stressful situation as
bullying. Several coping options are available to survive in a stressful situation such as changing
the situation, adapting to the situation, searching for more information or controlling one self
emotionally from reacting in an impulsive or retaliatory manner. It is argued that employees who
are constantly subjected to workplace bullying incidences are more likely to indulge in emotional
regulation by self coping which involves displaying fake emotions and hiding their genuinely felt
emotions producing high levels of emotional dissonance
40
Another theory which can shed light on how workplace bullying might create emotional
dissonance in employees is the theory of Façade creation which is one of the leading theoretical
pillars in the area of Person-organization fit and organization socialization literature. This theory
assumes that employees create fake images to appear more acceptable in the eyes of others and
give a signal that they have internalized the organizational values (Hewlin, 2003).
According to this theory when individual face inconsistency between their individual and
organizational values some employees may feel the need to hold back their own values and
pretend as if they have accepted organizational values (Hewlin, 2003). In the context of
workplace bullying this would mean that when employees are exposed to an intimidating
situation of workplace bullying their felt emotions of anger and retribution might come into
direct conflict with the required emotional displays and victims due to feeling in a helpless
situation might demonstrate fake emotions while dealing with the perpetrator of bullying creating
high levels of emotional dissonance in them. Thus, it is hypothesized:
Hypothesis 5: Workplace Bullying is positively related to Emotional Dissonance.
2.8 Mediation (Simple/Parallel)
2.8.1 Perceived Support types as mediators in the relationship of Workplace Bullying and Job
Strains (Job stress, Job Burnout, Psychological Strain and Intensions to quit).
Recent research has started to investigate perceived organizational support's role as mediating
mechanism between positive treatment and individual outcomes. It is asserted that even though
previous studies on perceived organizational support have identified the determinants and
outcomes, there is a strong need by future studies to focus on the pathway running from POS to
outcomes and the underline theories which support such arguments (Rhoades & Eisenberger,
41
2002). It is the contention of the present study that perceived organizational support is an
important intervening process through which the effect of workplace bullying might be revealed
on employee’s outcomes.
Past research examining the organizational climate and bullying relationship indicate that
although role stress did not account for substantial variations in workplace bullying whereas
exposure to bullying at work was a significant predictor of employee’s perceptions of role stress
(Giorgi & Majer 2009; Hauge, Skogstad & Einarsen, 2011). Hauge et.al. (2011) found
significant negative relationships between exposure to bullying and organizational climate. In
line with this research it is contended that individuals who are exposed to bullying develop
negative cognitive evaluations about the work environment particularly these tortured individuals
feel that the organization does not value them, consider them as a part of the organization and
does not care for their well-being. Their negative assessments about organization’s support might
then translate into higher levels of stress and burnout. Since victims of bullying behaviors might
develop negative interpretations about organization’s support, if these negative actions are not
lessened they may lead to long-term harmful effects for the victims such as higher stress
complaints and burnout.
The conservation of resource model (Hobfoll, 2010) postulates that resource loss is taxing for an
individual, the individuals who have lost their resources as a result of encountering stressor may
invest other resources to compensate for the net loss (Pearlin et al., 1981). Replenishment is the
most explicit means by which resources are gained. However, when individuals face threatening
situations such as workplace bullying they might lose resources which might lead to lower
perceptions of organizational support.
42
In the long term, exposure to repetitive and persistent stressful incidences such as workplace
bullying might leave individuals with fewer resources as they develop negative perceptions and
cognitive interpretations about the organization. These individuals with lower perceptions of
organization support might become incapable to endure more threats to resource loss creating
high levels of stress and burnout. According to Hobfoll (2010), resource loss spirals are more
important and go about more rapidly than resource gains. Therefore, it is contended that bullying
might lead to higher levels of stress and burnout through a vicious and contagious effect of
continuous resource loss through lower perceptions of organizational support. Thus, it is
hypothesized:
Hypothesis 6a: Perceived Organization Support mediates the relationship of workplace bullying
with job strains (Job stress, Job burnout, Psychological Strain and Intensions to quit).
Hypothesis 6b: Perceived Coworker Support mediates the relationship of workplace bullying
with job strains (Job stress, Job burnout, Psychological Strain and Intensions to quit).
Hypothesis 6c: Perceived Supervisor Support mediates the relationship of workplace bullying
with job strains (Job stress, Job burnout, Psychological Strain and Intensions to quit).
2.8.2 Mediation Effects of Emotional Dissonance in Relationship of Workplace Bullying and
Job Strains (Job Stress, Job Burnout, Psychological Strain and Intensions to quit)
The COR theory (Hobfoll, 1989; 1998) adds important insights into how workplace bullying
incidences leads to stress and job burnout through the mechanism of emotional dissonance. The
COR theory argues that individuals seek to obtain and sustain their resources. Hobfoll (1989,
2002) pointed out different kinds of resources (things, situations, individual traits and energies).
43
Stress occurs in conditions where there is a risk of decrease of resources or inability to increase
one's resources. Moreover, stress emerges from persistent and demanding situations of the work
that necessitate individuals to invest energies so as to survive in them and specifically from the
kinds of constant work burdens and resource discrepancies (Cordes & Dougherty, 1993; Hobfoll,
1989, 2002).
It is argued that workplace bullying is an acute stressor faced by employees which results in
emotional and psychological resource loss. In order to cope with such an emotionally demanding
situation of workplace bullying employees expend more emotional energies by altering the
emotions according to the organizational requirements so that they can meet their work demands
and keep their professional relationships intact which creates emotional dissonance as the
emotions which they are displaying do not truly reveal their inner felt natural emotions which
result in losing more emotional resources. Further, this resource loss of an emotionally dissonant
employee would further become evident in the form of burnout and stress as emotionally
dissonant employees shows inconsistent portrayed and felt emotions which exasperates the
vicious cycle of resource loss creating high levels of stress and burnout.
It has been corroborated in the past research on workplace bullying that negative behaviors at the
workplace creates feelings of isolation, estrangement and impaired identity in employees.
Moreover, it has also been viewed in the literature on emotional dissonance that weak
identification with the organization leads to emotional dissonance as these employees do not
internalize the organizational roles genuinely and consider the performance of these roles as a
threat to their identity and express fake emotions promoting emotional dissonance (Ashforth &
Humphrey, 1993; Elsbach & Bhattacharya, 2001; Mishra & Bhatnagar, 2010).
44
Bridging the literatures on bullying and emotional dissonance, it is contended that when
employees are subjected to negative treatment at work they feel emotionally dismembered and
hurt which creates emotional dissonance in them as these employees feel betrayed and do not
truly reveal the inner negative emotions felt by them and display emotions only which are
required by the organizational norms thus creating emotional dissonance which further creates
stress and burnout.
Another useful theory which can explain the mediational mechanisms of emotional dissonance
between workplace bullying and outcomes is the cognitive theory of stress and coping. This
theory contends that stress is created as a consequence of the combined effects of both the person
and situation which is evaluated by the individual as a demanding or a situation which goes
beyond one’s resources and threaten well-being (Lazarus & Folkman, 1984b). If the individual
evaluates a demanding situation as stressful primary appraisal occurs. After the cognitive
assessment of one’s demanding situation one tries to establish coping strategies to avert or at
least survive in the situation.
Moreover, existing research on the relationship between stressful and demanding situations and
individual outcomes such as psychological health complaints are mediated by coping
mechanisms. This line of inquiry stresses how an individual manages single or many stressful
situations (Billings & Moos, 1984; Folkman, Lazarus, Gruen & DeLongis, 1986).
It has been established in the literature on coping that individuals who are constantly exposed to
uncontrollable situations are more vulnerable to negative treatment, become highly passive in
their coping strategies and eventually suffer from health complaints such as stress, anxiety and
depression (Folkman, Lazarus, Gruen & DeLongis, 1986). The literature on coping depict that
45
single or isolated events are not likely to affect an individual’s health and well-being whereas
situations which are repeatedly threatening or intimidating for individuals have long term
consequences for individuals and these people might use maladaptive and poor coping strategies
(Krohne & Laux, 1982; Schultz & Moore, 1984; Solano, Batten & Parish, 1982). Applying this
argument in the context of workplace bullying, it is believed that when individuals face
workplace bullying leads to stress and burnout through its effects on emotional dissonance.
When employees encounter workplace bullying they engage in emotional regulation strategies
such as portraying fake emotions and suppressing true emotions to cope with such a stressful and
demanding situation on the job creating emotional dissonance in them. These emotionally
dissonant employees as a result of experiencing discrepant felt and expressed emotions generate
high levels of stress, burnout, psychological strain and intensions to quit. Thus, it is
hypothesized:
Hypothesis 7: Emotional Dissonance mediates the relationship of workplace bullying with Job
Strains (Job Stress, Job Burnout, Psychological Strain and Intensions to quit).
2.9 Sequential Mediation
2.9.1 Perceived Support types and Emotional dissonance as sequential mediators between
workplace bullying and job strains relationship
The Cognitive appraisal theory of stress considers two important mechanisms i.e cognitions and
coping as mediators between the stressors and strains relationship (Lazarus, 1966; Lazarus,
1991). According to this theory, a stressful event in the workplace leads to the activation of
cognitions called the primary appraisal process. Utilizing the above theory, it is asserted that
primary appraisal process occurs when employees are exposed to an intimidating situation such
as workplace bullying, these employees as a result of the abusive treatment subjectively and
46
cognitively develop lower perceptions of support of the supervisors/coworkers and of the
organization. The victims of bullying might cognitively interpret the workplace bullying as an
indication of lower support provided to them as they believe that the organization is not playing
its role to stop the bullying, supervisors are abusively using their authority to take advantage of
them and coworkers are ridiculing them or indulging in other kinds of humiliating treatment thus
generating lower perceptions of support.
The theory further assumes that once an individual who faces a stressful situation has cognitively
evaluated the situation as threatening to them the secondary appraisal process commences. The
Secondary appraisal process asserts that individuals decide what if anything can be done to stop
the harm (Smith & Lazarus, 1993). In this process, individuals determine their existing coping
options to deal with the stressor. Applying, the second assumption of this theory in the context of
workplace bullying, it is contended that secondary appraisal process begins when individuals
who interpret lower perceptions of support in response to workplace bullying indulge in
emotional self regulation creating higher levels of emotional dissonance. This means that the
targets as a result of viewing lower perceptions of support engage in faking their displayed
emotions to cope with a hostile and aggressive situation of bullying thus producing inconsistency
between their felt and displayed emotions resulting in emotional dissonance.
The Hobfoll’s conservation of resource theory is also a useful mechanism by which the
sequential role of Perceived support and emotional dissonance can be explained between the
workplace bullying and strains relationship (Hobfoll’s, 1988, 1989, 1998a). This theory asserts
that individuals struggle to obtain, increase and protect their values resources. Resources are
categorized as physical, psychological, social, emotional and work related aspects of one’s work.
This theory asserts that resource losses are more important than resource gains and highlights
47
two important principles of primary resource loss and secondary resource loss. Primary resource
loss means that whenever individuals face stressors at the workplace resource loss occurs which
creates strain.
In secondary resource loss, individuals try to regain their lost resources by investing additional
resources to offset the previous resources lost. However, the secondary resource loss assumption
views that if individual’s investment of additional resources further depletes one’s resources then
a resource spiral occurs creating strain and negative well-being for the individual (Hobfoll, 1988,
2002; 2010). Applying the theoretical foundations of this theory in this context, it is believed that
primary resource loss occurs when individuals who experience a stressor such as workplace
bullying creates psychological resource loss in the form of lower perceptions of support.
Past research has provided support that perceived organizational support is an important resource
(Hobfoll, 1989) which helps employees to deal with stressors. I believe that workplace bullying
as a stressor might result in the depletion of one’s psychological resources in the form of lower
perceived support as employees when faced with a hostile situation might lose psychological
resources and perceive that the organization, its representatives and the peers do not care about
them and do not value their well-being as well. Further, secondary resource loss occurs when
employees as a result of losing resources in the form of lower perceptions of support then put in
additional resources by faking their emotions and becoming more acceptable in front of
supervisors/coworker and subordinates which creates an inconsistency between their inner and
exhibited emotions resulting in higher levels of emotional dissonance. Since secondary resource
loss assumes that if investing additional resources further exhausts one’s resources then a vicious
cycle of resource depletion occurs producing higher levels of strain in the form of higher stress,
burnout, psychological strain and intensions to quit. Thus, it is hypothesized:
48
Hypothesis 8: Perceived support types (a) Perceptions of organization support (b) Perceptions of
supervisor support; and (c) Perceptions of coworker support and Emotional Dissonance
mediates the relationship of workplace bullying with Job Strains (Job Stress, Job Burnout,
Psychological Strain and Intensions to quit).
2.10 Moderation
2.10.1 The moderating role of CSE between workplace bullying and perceived support types
(POS, PSS and PCS).
As discussed previously CSE is a latent construct comprising of 4 key types which are self
efficacy, locus of control, self esteem and neuroticism. It is the assertion of the present study that
individuals who posses negative core self evaluation will strengthen the relationship between
workplace bullying and perceived organizational support.
Past research highlights that person who exhibit high neuroticism are highly negative and they
have a predisposition to interpret themselves, other people and the world in a negative manner
(Spector, Zapf, Chen & Frese, 2000). Spector et.al. (2000) propose that people who are more
neurotic may react more strongly to work related stressors than do individuals who are low on
neuroticism under the same situational circumstances. It is believed that since individuals with
negative core self evaluations view themselves and others negatively in the face of a stressor
such as workplace bullying might develop lower perceptions of organizational support.
Generalized self efficacy refers to a person's confidence in his/her capabilities to successfully
plan and implement a specific strategy to achieve goals (Bandura, 1996). Individuals low in self
49
efficacy perceive emotionally charged situations as more demanding than their high self efficacy
counterparts as the former types are hindered rather than challenged by novel, hard and changing
work and circumstances (Gist & Mitchell, 1992). Low self efficacious and low self esteem
individuals dislike themselves, do not believe in their capacity to achieve goals, give up in the
face of difficult situations and feel demotivated in case of negative feedback and failures than are
their high self efficacious counterparts (Bandura, 1996). It is asserted that when individuals with
negative core self evaluations encounter an unfavourable situation such as bullying might
develop lower perceptions of organizational support since these individuals might feel highly
demoralized and feel that the organization does not care about them and does not value them.
Past research analyzing the trait of locus of control highlights that individuals who posses an
internal locus of control since hold themselves accountable for their lives, these individuals will
be able to respond to stressful and intimidating situations in a calm and relaxed manner (Glass &
Singer, 1972). In contrast, individuals possessing an external locus of control feel a lack of
control in their lives in case of unfavorable work situations (Schonpflug, 1983). It is argued that
since individuals with a negative core self evaluation have an external locus of control, they will
blame external factors and luck when they will be faced with an intimidating situation such as
workplace bullying which might lead to lower perceptions of organization support. In the face of
bullying, individuals with negative core self evaluation will not make any effort to control the
instances of bullying and let the bullying to occur unfettered which might ultimately create lower
perceptions of organizational support in them.
Dodgson & Wood (1998) asserted that individuals who counter setbacks with flexibility may do
so because they emphasize on more favorable thoughts about themselves. People with negative
CSE may be less likely to recover from setbacks as they are more likely to highlight their
50
negatives and underscore their positives (Bono & Colbert, 2005). Therefore, in the light of above
arguments, it is the assertion of the present study that since individuals who possess negative
core self evaluations consider themselves worthless, incompetent and lack control in the face of
an intimidating situation such as workplace bullying might develop lower perceptions of
organizational support because these individuals do not work harder to lessen instances of
bullying, interpret the organization as highly unsupportive, become extremely demoralized and
view the organization favoring the bully which will lead to lower perceptions of organizational
support. Thus, it is hypothesized:
Hypothesis 9a: Core Self Evaluations act as a moderator between Workplace Bullying and
Perceived Organization Support such that this negative relationship is stronger in case of
negative core self evaluations.
Hypothesis 9b: Core Self Evaluations act as a moderator between Workplace Bullying and
Perceived Coworker Support such that this negative relationship is stronger in case of negative
core self evaluations.
Hypothesis 9c: Core Self Evaluations act as a moderator between Workplace Bullying and
Perceived Supervisor Support such that this negative relationship is stronger in case of negative
core self evaluations.
2.10.2 Core Self Evaluation as a moderator between workplace bullying and emotional
dissonance
Individuals who posses negative core self evaluation tend to evaluate themselves and others in a
constantly negative manner across different situations. These individuals view themselves as
worthless, incompetent, lack control of their lives and are emotionally unstable (Judge et al.,
51
2004; Judge 2007). It is the assertion of the present study that individuals with negative CSE
might strengthen the workplace bullying and emotional dissonance relationship whereas people
with positive CSE might weaken this relationship.
It has been contended in past research that people who have higher self efficacy are less likely to
face emotional dissonance (Heuven, Bakker, Schaufeli, Huisman, 2006). Higher self efficacy
persons are reported to be less nervous and aggravated and suffer less from anxiety prone
circumstances (Bandura, 1977; 1986). Moreover, highly self efficacious individuals are more
likely to employ diverse range of useful survival tactics than people having lower self efficacy
(Lazarus & Folkman, 1984). On the other hand, individuals low in self efficacy might feel more
anxious and tense and employ maladaptive coping strategies in the face of stressful situations.
It has also been established in past research that employees who have low self esteem, have a
negative evaluation about themselves and who dislike themselves are more likely to resist
organizational display rules requiring them to show positive emotions (Abraham, 1999).
Moreover, it is also contended that employees who have negative core self evaluations are more
likely to form opinion based on what others think about them, are always looking for outside
approval and since they do not believe in themselves they are always trying to make others
happy. It is the believed that since such people with negative CSE are unable to handle tense and
stressful work situations, dislike themselves and do not have confidence in their abilities when
faced with such bullying experiences are more likely to pretend, show fake emotions and hide
their real emotions in front of the bully resulting in high levels of emotional dissonance.
Previous research on emotional regulation highlights that employee’s lower self efficacy
strengthened the emotional job requirements and emotional dissonance relationship (Heuven,
52
Bakker, Schaufeli & Huisman, 2006). Therefore, it is asserted that in the face of a stressful
situation such as bullying and negative behaviors at work, employees with negative core self
evaluations may not invest more resources to regulate negative emotions and thus face higher
emotional dissonance. Since individuals with negative core self evaluations might feel more
anxious and insecure in the face of stressful and intimidating situation such as workplace
bullying; these individuals might feel difficulty in adjusting themselves to portray organizational
desired emotions thus there might be higher discrepancy for these individuals expressed and felt
emotions creating emotional dissonance.
It has been established in the literature on personality that individual who have positive core self
evaluations are highly effective when dealing with negative feedback as these individuals might
not show strong negative emotional responses in case of a setback. Whereas individuals with
negative core self evaluations might use ineffective or maladaptive coping strategies when
dealing with unfavorable situations as these individuals might elicit strong negative emotions in
response to setbacks (Brown & Dutton, 1995; Kernis, Brockner & Frankel, 1989).
Extending this argument to workplace bullying, it is asserted that people with negative CSE are
unable to cope in the face of workplace bullying as they will elicit emotions of anger, frustration
and anxiety and might feel the incapability to undo the harmful effects of bullying which might
create higher levels of emotional dissonance. Individuals who posses negative core self
evaluations since exhibit low self esteem and efficacy might feel highly pessimistic and anxious
and lack of control when dealing with a highly unfavorable situation such as bullying which
might create a higher incompatibility between their inner and expressed emotions depicting
emotional dissonance. Thus, it is hypothesized:
53
Hypothesis 10: Core Self Evaluations act as a moderator between Workplace Bullying and
Emotional Dissonance such that this positive relationship is stronger in case of negative core
self evaluations.
54
CHAPTER 3
RESEARCH METHODOLOGY
Chapter Overview
The present chapter gives information regarding the research methods used to conduct the
current thesis. This chapter informs the reader of the design of the research, sampling techniques
data collection procedures, study measures, Data Analytical techniques, Sample characteristics
and findings pertaining to confirmatory factor analysis of the measured items that capture the
underlying study constructs.
3.1 Research Design
A study's research design provides a broader layout of how research will be conducted.
According to the research objectives and research model this study examines how workplace
bullying as a stressor affects job strains through the parallel and sequential mechanisms of
Perceived Support Types (POS, PSS and PCS) and Emotional dissonance. It also examines the
moderating role of CSE personality trait between workplace bullying and Emotional Dissonance
and Perceived Support Types. Therefore, this is a quantitative research study (Hypothesis
testing) in which the above proposed relationships are tested through the survey method.
Longitudinal research design with temporally separated data at three different time periods was
used to collect data on the proposed model variables. This research was done in pure longitudinal
55
design (West, Aiken & Todd, 2002). In a recent meta-analysis on the outcomes of workplace
bullying, it was corroborated by researchers that quantitative studies conducted on workplace
bullying shows that majority of the quantitative studies employed cross-sectional designs and
fewer than 15 studies have been conducted that have employed longitudinal research designs
(Nielsen & Einarsen, 2012). Moreover, due to inconsistent and weak relationships between
workplace bullying and outcomes, greater longitudinal designs are required which can highlight
the long-term consequences of how workplace bullying shows its effects on outcomes (Vie,
Glaso & Einarsen, 2011). In addition, since casual relationships take time to unfold therefore
future researches can increase value in the workplace bullying research by employing
longitudinal research designs (Nielsen & Einarsen, 2012). Therefore, following the
recommendations of the recent quantitative meta-analysis on workplace bullying (Nielsen &
Einarsen, 2012), longitudinal research design was adopted to test the proposed research model.
The researchers studying methodological issues in mediational models in behavioral research
also highlight that cross-sectional research designs are inadequate and depict an imperfect picture
of the reality (Ployhart & Vandenberg, 2010; Selig & Preacher, 2009). Cross-sectional research
designs where data on the constructs is collected at a single time period poses serious problems
of causality of correlated variables. Thus, temporally separating variables at different
measurement points addresses the reverse causality issues arising because of single shot
collection of data.
Therefore in order to fully identify and explain the cause and effect relationships particularly the
underlying mechanisms of a phenomenon, it has been suggested that future research on
workplace bullying should employ longitudinal research designs with appropriate time lag
between different measurement points (Ployhart &Vandenberg, 2010; Nielsen & Einarsen,
56
2012). Therefore, in order to measure the proposed mediating (both parallel and sequential)
mechanisms of workplace bullying on job strains through Emotional dissonance and Perceived
Support types the present study employed longitudinal research design with 3 wave temporally
separated data.
Another concern regarding the longitudinal research designs is the appropriate time lag between
the different survey points as it has been established in previous research on workplace bullying
that both short time gaps as well as long time gaps can be useful in providing insights about the
causal mechanisms of the consequences of workplace bullying (Nielsen & Einarsen, 2012).
Nielsen & Einarsen (2012) in their meta-analysis highlight that studies of workplace bullying
which have employed longitudinal research designs have used 2 or multiple measurement points
for tapping the model variables. The minimum recommended time lag between different time
periods according to these studies on workplace bullying is approx 1 to 3 months. In the present
study, longitudinal design with temporally separated data was followed by which data was
collected at 3 different time intervals, i.e. time 1, time 2 and time 3 with a time interval of approx
3 months. Bullying was tapped at time 1 and Core Self Evaluation was tapped at time and time 2.
The current thesis mediators (i.e. Emotional Dissonance, Perceived Organization Support,
Perceived supervisor support and perceived co-worker support) were tapped at time 2 and time 3
and outcomes of Job Stress, Job Burnout, Psychological Strain and Turnover Intensions were
tapped at time 3.
3.2 Population
57
To capture maximum variation in the research model variables, data was collected from diverse
occupational groups and organizational settings. The identified population for the current study
were employees working in the service sector specifically three major industry/sectors i.e.
Telecom, banking and educational sector were the target population. Past studies on workplace
bullying highlight that this topic has more relevance in the service sector (Hutchinson et. al.,
2010; Monks, Smith, Naylor et al., 2009; Matthiesen & Einarsen., 2004).
The sample respondents from these target sectors belonged to different managerial levels in the
organization. Data was collected from 15 organizations out of which they were 3 telecom
organizations, 6 organizations belonged to the banking sector and 6 were from educational
Sector. The surveys were distributed to employees in 6 banks which comprised of 5 local banks
and 1 was a multinational bank. The 3 telecom sector organizations consisted of two
multinational telecom companies and 1 was a semi-government telecom company. The data
collected from educational sector comprised of 6 different organizations out of which 3 were
leading universities in Islamabad and 3 were colleges.
3.3 Sampling Technique and Sample Size
Sample size depends on the type of research design. Bartlett, Kotrlik & Higgins (2001) suggested
that minimum size should be 200 and for moderately complex research models should be around
400 (Boomsma, 1983). Since the present study employs longitudinal design with temporally
separated data at three different time periods therefore the researcher was able to achieve a
sample of above 300 (details explained the next section) for the current study which can be
considered large enough sample size when using a longitudinal research design with multiple
58
time periods (Boomsma, 1983). The sampling strategy used to collect data and manage the
questionnaire was non-probability convenient sampling.
3.4 Data Collection Procedures and Sample
Data was collected through questionnaires that were personally administered at three different
time periods i.e. time 1, time 2 and time 3 from the same respondent. The surveys were
circulated to employees working in the offices of the 15 organizations belonging to banking,
telecom and educational sector situated in the twin cities of Rawalpindi and Islamabad in
Pakistan. A cover letter explaining the scope and objective of the current research study was
provided to assure respondents of confidentiality. Participation in the study was voluntary.
After the cover letter, the first page of the questionnaire highlighted a Questionnaire ID and the
time-1 date and respondent name/identifying information such as employee no or any unique ID
which was used to track the same respondent at time-2 and time-3 after gaps of approx 3 months.
The time-2 and time-3 questionnaire also had QID and time interval date to appropriately
measure the time gap and follow up on the same respondent successfully. The time-1
questionnaire consisted of the independent variable of workplace bullying measures. The
moderator variable of core self evaluation was measured at time-2 only and the mediator
variables of the study i.e. Emotional Dissonance, Perceived Organization Support, Perceived
supervisor support and perceived co-worker support were measured at time 2 and time 3. Finally,
the study's mediators and the outcomes i.e. Job stress, burnout, psychological strain and turnover
intensions were tapped at time-3.
At time-1, I distributed approx 700 questionnaires out of which 570 were completely usable
generating a response rate of 81% at this stage. After a time interval of three months I established
59
communication with time 1 respondents and asked them to fill the time 2 questionnaire. At this
stage the complete useable time-1 and time-2 questionnaire for the same respondent were 455
resulting in a response rate of 65%. After time lag of 3 months from the date on which the time-2
questionnaire was distributed, I approached the same respondents of time-1 and time-2 and
requested them to fill out the third wave questionnaire comprising the mediator and outcome
variables of the study. At the third time period the complete matched time 1, time 2 and time 3
questionnaire for the same respondent resulted in a final sample size of 334, thus making the
final response rate for all three time periods of 48%.
The sample characterized employees belonging to a number of different occupations including
administrative staff, accountants, bankers, junior managers, supervisors, marketing and HR
specialists, academicians and top managerial employees. Participants in the survey had a diverse
background and belonged to different management levels ranging from lower management levels
to top management. Males constituted 61% and females were 39% of the total population. Most
of the participants belonged to the Human resource management division (37.5%), had Masters
degree as their qualification (78.7%) and were on the junior management levels (52.7%). In
addition, the frequencies revealed that bulk of the sample consisted of employees who had
accounting and finance as their field of specialization (37.4%). The sample statistics revealed
that the employee's average working experience in the current organization was 3.42 (SD=4.41)
and total working experience was 6.54 (SD=7.13) years. The average respondent age was 30.1
years (SD=8.58).
3.5 Measures
60
English is the official language of communication used at work and is also used as a medium of
instruction at college and university level. Majority of the respondents were Masters degree
holders indicating their ability to understand and interpret statements in English language. The
study's sample respondents comprised of mostly entry level managers which also highlights that
language might not have posed a serious threat in understanding the questionnaire. Moreover, for
entry level management positions, even for front desk positions, the hiring and selection
procedures particularly from the sectors I collected data i.e. banks, telecom and educational
sector requires the candidates to exhibit adequate proficiency in the speaking and reading of
English language therefore language was not a barrier in understanding the questionnaire.
Furthermore, I employed well-established and well-validated scales for measuring variables
which have been tested in a number of countries across the globe which are simple to
understand. In addition, past studies published in mainstream journals (Abbas, Raja, Darr &
Bouckenooghe, 2012; Butt, Choi & Jaeger, 2005; Raja, Johns & Ntalianis, 2004) have utilized
English questionnaires in Pakistan. Lastly, the researcher personally administered the
questionnaires to the respondents. Moreover, due to time lagged data the procedures were
explained on how to fill the questionnaire and brief discussion with the respondent indicated that
they were equipped to use and apply English as medium of communication. Therefore, on the
basis of the above justifications, I did not translate my questionnaires into Urdu.
3.5.1 Workplace Bullying
61
The most commonly used method to tap bullying at the workplace is through the survey
technique where members highlight how often and repeatedly they are subjected to bullying
behaviors at the workplace or unfavorable actions that do not directly ask them to identify the
concept as well as accept workplace bullying behaviors. The Negative acts questionnaire revised
(NAQ-R) is the most widely employed scale to measure workplace bullying whose psychometric
strength and superiority has been methodically examined and established (Einarsen, Hoel &
Notelaers, 2009).
Workplace bullying was measured at time-1 by the Negative acts questionnaire (NAQ)
formulated by Einarsen and his associates (Einarsen & Hoel, 2001; Einarsen, Hoel, Zapf &
Cooper, 2003). The NAQ questionnaire consists of eight items (NAQ; Einarsen & Raknes,
1997).The NAQ scale which measures bullying behaviors at the workplace has been translated
into different languages and well-validated and tested across a number of countries across the
globe (Einarsen & Skogstad, 1996; Hoel, Cooper & Faragher, 2001; Keashly & Jagatic, 2000;
Power et. al., 2013). Participants who filled up the surveys were asked to highlight how
regularly they had been subjected to such negative behaviors at the workplace for the past six
months time period. Sample items are “I have been exposed to an impossible workload to carry
out.”; “I have been ridiculed in relation to my work”; and "I have been controlled so much that I
feel extremely insecure". The sample respondent reported their answers on a 7 point likert scale
where 1=Never, 2=Once in a while, 3=Once in six months, 4= Once in a month, 5= After every
two weeks, 6=Every week and 7=Daily. Past studies have also demonstrated high construct
validity and reliability for this scale (Einarsen & Raknes, 1997). Higher scores reflected higher
frequencies of perceptions of workplace bullying.
62
In my study, alpha reliability for this measure was α= 0.88. In order to test validity for this
measure CFA was run as full measurement model in SEM. The full measurement model was run
as a first step of Structural Equation Modeling which showed that workplace bullying measure
demonstrated adequate discriminant validity with other constructs as well as convergent validity
with its own construct. Moreover, all workplace bullying items demonstrated acceptable factor
loading above 0.4 (see Figure 2 for details of factor loadings). The convergent validity tests also
indicate that workplace bullying had composite reliability (CR) of 0.87 which is above the
minimum required of 0.6 and the Average variance extracted score (AVE) of 0.50 demonstrating
adequate convergent validity (see Full measurement model section in SEM).
3.5.2 Core Self Evaluations (CSE)
Core self evaluations items were tapped at time 2 utilizing a twelve-item scale (Judge, Erez,
Bono & Thoresen, 2002). Sample items are “I am confident 1 get the success I deserve in life”,
"I determine what will happen in my life"; and "I am capable of coping with most of my
problems". The respondents answered on a 7 point scale which were 1=Strongly Disagree,
2=Disagree, 3=Moderately Disagree, 4=Neutral, 5=Moderately Agree, 6=Agree and 7=Strongly
Agree. Higher scores on the scale indicated individuals possessing positive CSE whereas lower
scores depicted individuals having a negative CSE predisposition.
Core self evaluation was measured at time -2 along with the mediators. Tests of reliability and
validity were conducted to determine acceptability of core self evaluation measure in this study.
This scale in the present study had Cronbach alpha reliability of α= 0.81. The findings of the full
measurement model depict that core self evaluations items had acceptable factor loadings,
demonstrated adequate discriminant validity and convergent validity (see Full measurement
model section in SEM) (see figure 2 for details of these validity tests).
63
3.5.3 Perceived Organizational Support
In the present study, perceptions of organizational support was tapped at time 2 and time 3 by a
shortened version of POS devised and developed by Eisenberger et.al., (1986). Examples of the
items were “My organization cares about my opinions’’; ‘‘Help is available from my
organization when I have a problem” and "My organization would forgive an honest mistake on
my part." The respondents reported their answers on a likert scale which had anchors ranging
from 1=Strongly disagree to 7=Strongly agree. Higher scores reflected higher perceived
organizational support.
Tests of reliability revealed that deleting two items from the measures of Perceived Organization
support improved the scale reliability from .82 to .89. The findings of the Confirmatory factor
analysis (Full CFA/Measurement model) also highlighted that these two items had low factor
loadings i.e. below the 0.4 acceptable limit. Thus, I decided to remove these two items from the
scale of Perceived Organization Support. The removed items were "If given the opportunity, my
organization would take advantage of me" (reverse coded) and " My organization shows little
concern for me (reverse coded)."
Since Perceived Organization Support was measured at two different time periods along with
other mediators, it was necessary to conduct a series of Confirmatory Factor Analysis for
Perceived Organization Support scale both at time-2 and time-3 to establish the discriminant
validity of this scale. Different models of CFA consisting of alternate pairing of perceived
organization support scale with other variables were run i.e 4 factor models, 3 factor and 2 factor
64
paired CFA comparisons established Perceived Organization support's discriminant validity (see
Table 2 and 3 for CFA results model fit statistics).
The tests of validity (discriminant and convergent) clearly showed that Perceived Organization
Support scale had adequate acceptability in this study. The rest of the Perceived Organization
scale items after deleting the two reverse coded items also displayed adequate factor loadings as
well. The Composite reliability (CR) of 0.89 and the Average variance extracted (AVE) of 0.58
showed that this scale displayed convergent validity as well. Since, Perceived Organization
Support was also measured both at time-2 and time-3 in this study, a test-retest analysis of
reliability was conducted to establish the consistency in the reliability of Perceived Organization
Support at both time periods. The Cronbach alpha, for Perceived Organization Support measures
at time-2 was α= 0.89 and at time-3 was α= 0.90 (refer to Table-1).
3.5.4 Perceived Supervisor Support
Perceptions of supervisor support was measured at time-2 and time-3 by a seven item scale
devised by Tsui, Pearce, Porter & Tripoli (1997). Sample items were “My supervisor seems
willing to listen to my problems”, " My supervisor is considerate of subordinate’s feelings" and "
My supervisor is friendly". Respondents rated on a seven point scale i.e. ranging from Strongly
disagree=1 to Strongly disagree=7. Higher responses on this scale indicated increased levels of
perceptions of supervisory support.
The reliability test indicated that deleting two items from the measures of Perceived Supervisor
support improved the reliability of the scale from .83 to .89. The results of CFA model (Full
measurement model) also highlighted that these two items had weak factor loadings i.e. less than
65
the acceptable limit of 0.4. As a result I in turn removed these two items from the scale of
Perceived Supervisor Support. The removed items were " I don’t really feel that my supervisor
and I are working towards shared or team objectives (reverse coded)" and "My supervisor seems
to be rather distant and unapproachable (reverse coded)."
A number of Confirmatory Factor Analysis were performed for perceptions of supervisor support
at both time 2 and time 3 to establish the discriminant validity of this scale with other variables
measured from the same source and within the same time period. Different models of CFA
consisting of alternate pairing of perceived supervisor support scale with other variables were
run i.e. 4 factor models, 3 factor and 2 factor paired CFA comparisons established Perceived
Supervisor support's discriminant validity (see Table 2 and 3 for CFA results model fit statistics).
The tests of convergent validity clearly showed that Perceived Supervisor Support scale had
adequate acceptability in this study. The rest of the Perceived Supervisor Support scale items
after deleting the two reverse coded items also displayed adequate factor loadings as well (see
figure 2). The results of CR of 0.89 and the AVE of 0.62 showed that this scale displayed
acceptable convergent validity as well. Perceived Supervisor Support along with the other
mediators was also measured both at time 2 and time 3 in the current study, thus the results of the
test-retest reliability indicated higher and consistent reliability of Perceived Supervisor Support
at both time periods. The cronbach alpha, for Perceived Supervisor Support measures at time-2
was α= 0.89 and at time-3 was α= 0.90 (see Table 1).
3.5.5 Perceived Coworker Support
66
Perceived Coworker support was also measured along with the other mediators at time-2 and
time-3 as well. Perceived Co-worker scale was tapped by a 3 item scale developed by Staw,
Sutton & Pelled (1994) perceived co-worker support scale. Items of this scale are “My
coworkers give me the help I need to do my job”; " I and my coworkers share news about
important things that happen at the organization" and " I and my coworkers stick together".
Responses to the scale were tapped through a seven point scale ranging from 1= strongly
disagree to 7= strongly agree. Higher scores reflected higher perceived co-worker support. In the
current study, the alpha reliability for this scale is α= 0.84.
A number of Confirmatory factor Analysis were run for Perceived Co-worker scale both at time-
2 and time-3 to establish the discriminant validity of this scale. Different models consisting of
alternate pairing of perceived co-worker support scale with other variables were run i.e. 4 factor
models, 3 factor and 2 factor paired CFA comparisons corroborated Perceived Co-worker
support's discriminant validity at both time-2 and time-3 (see Table 2 and 3 for results of model
fit statistics). The results of the full measurement CFA also showed adequate factor loadings of
the items of this scale. The composite reliability (CR) of 0.81 and the average variance extracted
(AVE) of 0.60 showed that this scale displayed convergent validity as well. The results of the
test-retest reliability depicted that Perceived co-worker support demonstrated higher and
consistent reliability for both time periods i.e. the reliability for PCS at time-2 was α= 0.84 and at
time-3 was α= 0.85 (see Table 1).
3.5.6 Emotional Dissonance
67
Emotional dissonance Items were also measured at time-2 and time-3 using a five item scale
devised by Zapf, Mertini, Seifert, Vogt, Isic & Fischbach (2000). Respondents were asked to
consider their job and indicate how often did they feel this way in their job. Sample items were
“During your work, how often do you have to suppress your own feelings (e.g., irritation) to give
a ‘neutral’ impression?”, "During your work, how often are you unable to show your
spontaneous feelings?" and " During your work, how often do you have to express positive
feelings towards others while you actually feel indifferent?." Responses were measured with a
seven point scale with anchor points of 1=Never, 2=once in a while, 3= rarely, 4=sometimes,
5=often, 6=usually and 7=always. Higher scores reflected higher reported Emotional
Dissonance. The Cronbach alpha for Emotional Dissonance at time-2 in this study was α= 0.84.
A number of Confirmatory Factor Analysis were run for Emotional Dissonance scale both at
time-2 and time-3 to establish the discriminant validity of this scale. Different models consisting
of alternate pairing of Emotional Dissonance scale with other variables were run i.e 4 factor
models, 3 factor and 2 factor paired CFA comparisons corroborated Emotional Dissonance's
discriminant validity at both time-2 and time-3 (see Table 2 and 3 for results of model fit
statistics). The results of the full measurement CFA also showed adequate factor loadings of the
items of this scale. The CR (Composite reliability) of 0.83 and the AVE (Average variance
extracted) equivalent to 0.5 showed that this scale displayed convergent validity as well. The
results of the test-retest reliability depicted that Emotional dissonance demonstrated higher and
consistent reliability for both time periods i.e. the reliability for Emotional Dissonance at time-2
was α= 0.84 and at time-3 was α= 0.86 (see Table 1).
3.5.7 Job Stress
68
Job stress items were measured at time-3 using a 13 item scale (Parker & Decotiis, 1983).
Sample items included “My job gets to me more than it should”; “There are lots of times when
my job drives me right up the wall” and " I have felt fidgety or nervous as a result of my job".
Responses were measured with a seven point scale with anchors ranging from Strongly disagree
(1) to Strongly agree (7). The mean of the scale was calculated so that higher responses signaled
higher reported job stress. The cronbach alpha reliability for this scale in this study was α= 0.92.
Since Job Stress was measured at time-3 with other outcomes of this study, therefore it was
essential to perform Confirmatory Factor Analysis for Job Stress scale at time-3 to establish the
discriminant validity of this scale. Different models of CFA consisting of alternate pairing of job
stress scale with other variables were run i.e. 4 factor models, 3 factor and 2 factor paired CFA
comparisons established Perceived Organization support's discriminant validity (see Table 2 and
3 for CFA results model fit statistics). The tests of validity (discriminant and convergent) clearly
showed that Job stress scale had adequate acceptability in this study. The job stress scale items
displayed adequate factor loadings as well. The CR of 0.92 and the AVE of 0.50 showed that
this scale displayed convergent validity as well.
3.5.8 Job Burnout
Job burnout was measured at time 3 by a seven item scale formulated by Pines & Aronson,
(1988). Sample items for this scale were “I feel emotionally exhausted ”; "I feel depressed", "I
feel troubled." Respondents reported their answers on a seven point scale ranging from 1=Never,
2=once in a while, 3=rarely, 4=sometimes, 5=often, 6=usually and 7=always. Higher responses
on this scale reflected higher job burnout. The Cronbach alpha for this measure was α= 0.92.
69
Different models of CFA consisting of alternate pairing of job burnout scale with other variables
were run i.e. 4 factor models, 3 factor and 2 factor paired CFA comparisons established Job
burnout's discriminant validity (see Table 2 and 3 for CFA results model fit statistics). The tests
of validity (discriminant and convergent) clearly showed that Job burnout scale had adequate
acceptability in this study. The job burnout scale items displayed adequate factor loadings as
well (see Figure 2). The CR of 0.92 and the AVE of 0.63 showed that this scale displayed
convergent validity as well.
3.5.9 Psychological Strain
Psychological Strain was measured at time-3 adopted from the twelve item scale of General
Health questionnaire (GHQ) (Bank, Clegg, Jackson, Kemp, Stafford & Wall, 1980). The survey
participants were asked to highlight the degree to which they have been experiencing a number
of emotional and psychological symptoms over the past few weeks. Sample items are “I have
lost much sleep over worry”; " I have not been able to concentrate on whatever I’m doing"; and "
I feel constantly under a strain". Answers were tapped on a 7 point scale ranging from 1=Never
to 7=Always. Higher rates on this scale indicated higher psychological strain. The alpha
reliability for this scale was α= 0.94.
Different models of CFA consisting of alternate pairing of psychological strain scale with other
variables were run i.e. 4 factor models, 3 factor and 2 factor paired CFA comparisons established
Job burnout's discriminant validity (see Table 2 and 3 for CFA results model fit statistics). The
tests of validity (discriminant and convergent) clearly showed that psychological strain scale had
adequate acceptability in this study. The psychological strain items displayed adequate factor
loadings as well. The composite reliability of 0.95 and the average variance extracted of 0.59
showed that this scale displayed convergent validity as well.
70
3.5.10 Intensions to quit
Intensions to quit was tapped at time 3 utilizing a three item scale adopted from Cammanan,
Fichman, Jenkins & Klesh (1982). Examples of the items are “It is highly likely that I will look
for a new job in the next year"; and " I often think about quitting/ I often think about leaving the
organization". Respondents reported their answers on a 7 point scale which ranged from Strongly
disagree (1) to Strongly agree (7). Higher values on the scale indicated higher intensions to quit.
The results of the reliability analysis revealed that removing one item from the intensions to quit
scale improved the scale reliability from 0.66 to 0.82. The confirmatory factor analysis results
(refer to the full measurement model) indicated that this item had low factor loadings which were
less than the standard limit. Thus, I decided to remove this item from the intensions to quit scale.
The deleted item was " If I may choose again, I will choose to work for the current organization
(reverse coded)." The alpha reliability for this scale after eliminating one item was α= 0.82.
Different models of CFA consisting of alternate pairing of psychological strain scale with other
variables were run i.e. 4 factor models, 3 factor and 2 factor paired CFA comparisons established
turnover intensions discriminant validity (see Table 2 and 3 for CFA results model fit statistics).
The tests of validity (discriminant and convergent) clearly showed that Turnover Intensions scale
had adequate acceptability in this study. The Turnover Intensions items displayed adequate
factor loadings as well (see Figure 2). The composite reliability (CR) of 0.85 and the average
variance extracted (AVE) of 0.75 showed that this scale displayed convergent validity as well.
71
Table 1- Test-Retest Reliability of variables measured at Time-2 and Time-3
Sr
No.
Variables Time-2 Time-3
Cronbach Alpha α
1. Emotional Dissonance .84 .86
2. Perceived Organizational Support .89 .90
3. Perceived Supervisor Support .89 .90
4. Perceived Co-worker Support .84 .85
3.6 Control Variables
Research in behavioral sciences depict that demographic factors like respondent's age, gender,
tenure, organization type, area of specialization are related to commonly studied dependant
variables in OB (Xie & Johns, 1995). These demographic variables were included in the
questionnaire at all three time periods and information on these variables were collected through
self-reports by the same respondent for time 1, time 2 and time 3. A one way Analysis of
variance revealed that only age and gender were significant for the study's outcome variables i.e.
job stress, turnover intensions, burnout and psychological strain. Thus, age and gender were
controlled for in all subsequent analysis.
72
3.7 Data Analysis Techniques
Data received at three different time periods i.e. at time 1, time 2 and time 3 from the same
responded was coded and entered into SPSS 20 software. Apart from the matched responses of
all three time periods another consideration while considering a complete response for all three
time periods was that there should be minimum missing values and those response cases where
there were large number of missing cases where omitted from the study. After data entry
frequencies were run to check for missing data and any errors. Missing values were treated and
errors were removed.
Reverse coded items were transformed back to be used in reliability analysis. Reliability analysis
for all study variables were conducted. Test-retest reliability was done of the variables which
were tapped more than a single time period. Means of the study variables were calculated i.e.
Workplace bullying tapped at time 1, Core Self Evaluation at time 2, Emotional Dissonance,
Perceived Organization support, perceived co-worker support and perceived supervisor support
tapped at time-2 and outcomes of Job Stress, Burnout, Psychological strain and turnover
intensions tapped at time-3. A number of confirmatory factor analysis (CFAs) were conducted
for the variables which were measured at a single point in time and reported from the same
source to ensure construct as well as discriminant validity among study variables at time 2 and
time 3. Descriptive statistics analysis were also performed. Corelational analysis was conducted
for the current study's variables and paired comparison correlations were run for the variables
tapped at time-2 and time-3 (see Table 4 and 5).
Hypothesis were tested and verified through two of the most recent, vigorous and highly
pertinent statistical techniques namely Structural Equation Modeling analysis (SEM) in AMOS
73
and Regression analysis conducted through the Process approach employing bootstrapping
technique using SPSS software. Firstly, Structural Equation Modeling analysis was applied to
run a full CFA model (called Full measurement model). On the basis of the credence in the CFA
model, structural models were run to confirm hypothesis. These structural path models were
performed to verify direct effects hypothesis. Then, parallel or simple mediation analysis were
performed by running structural paths for testing parallel mediation paths and verifying these
hypothesis and structural models for sequential mediation analysis was conducted to confirm the
sequential mediation hypothesis.
After verification of hypothesis through SEM technique, the present study also utilized Process
approach to test and confirm the hypotheses. Preacher and Hayes (2004) have provided model
templates which enables researchers and scholars to test their proposed research models
employing bootstrapping technique. Thus, I used model 4 (see Preacher & Hayes, 2004
templates) for testing simple mediation hypotheses. Then, according to these templates, Model
No. 6 was used to test and verify parallel and sequential mediation concurrently. Finally, I
utilized Model No.1 for testing the moderation hypotheses. On the basis of the significant
interaction terms for the outcomes, interaction plots are made to confirm the strength,
significance and direction of moderator at high and low values. Slope tests were also performed
and reported to reflect whether the moderator variable enhances/strengthens or buffers the
independent and dependant variable relationship at high or low values.
3.8 Confirmatory Factor Analysis
This study employed three wave longitudinal data for measuring the research model variables.
The variables were temporally separated at three different time periods to avoid reverse causality
74
issues and to ensure the effect of predictor on mediating and outcome variables was not plagued
by problems of single shot research designs. Therefore, even though collecting data at three time
period forms the unique strength of the study, I investigated the discriminant validity of the
variables by conducting a number of confirmatory factor analyses. According to the suggestions
offered by Anderson and Gerbing (1988; 1992), the current study compared a two factor model
with a one factor model for each of the probable combinations of variables which were tapped at
the same time period and from the same source.
Confirmatory factor analysis at time 2 comprised of 4 factor model which consisted of all four
mediators i.e. Emotional Dissonance, Perceived Organization Support, Perceived supervisory
support and perceived co-worker support vs. 1 factor, then 3 factor model indicating the
discriminant validity of the three Perceived Support types over a single factor model and then
finally 2 factor paired comparison model of each of the mediators with a single factor model.
The variables tapped at time 3 included the mediators and outcomes so CFAs were conducted at
time 3 for these variables as well. Firstly, Confirmatory factor analysis at time 3 comprised of 4
factor model comprising the study's mediator variables measured at time 3 i.e. Emotional
Dissonance, Perceived Organization Support, Perceived supervisor support and perceived co-
worker support vs. 1 factor, then 3 factor model indicating the discriminant validity of the three
Perceived Support types over a single factor model and then finally 2 factor paired comparison
model of each of the mediators with a single factor model. Moreover, the CFAs were run for all
the four outcome variables measured at time 3. These CFAs included 4 factor model of outcomes
i.e. Job Stress, Burnout, Psychological Strain and Turnover Intensions vs. 1 factor, then a 3
factor model comprising job strains were compared against a single factor and finally 2 factor
pairing of each of the outcome variables was done with a single factor model. The results of
75
these CFAs are presented in table 2 for time 2 variables and in Table 3 for time 3 variables which
show that in every case the unconstrained multiple factor desired model demonstrated superior
fit than the one factor constrained model.
3.8.1 Emotional Dissonance
Time 2 Analysis
A four factor model -factor model comprising Emotional Dissonance, Perceived Organization
Support, perceptions of supervisory and coworker support at time 2 (χ² =397.74, df=145, p≤
.001; CFI =.93 , GFI =.88 , NFI = .89 , RMSEA = .07) exhibited enhanced model fit than the one
factor model (χ² =1830.77, df=152, p<0.001; CFI = .59, GFI = .68, NFI= .53, RMSEA= .18)(see
Table 2).
The 2 factor paired comparison of Emotional dissonance with each of the other mediator
variables measured at time 2 revealed the superiority of each of the 2 factor models over the
single factor model establishing discriminant validity. The 2 factor model comprising Emotional
dissonance and Perceived Supervisor Support (χ² =82.89, df=31, p≤ .001; CFI =.97 , GFI =.95,
NFI = .95 , RMSEA = .07) demonstrated higher model fit than the one factor model (χ² = 213.82,
df=26, p<0.001; CFI= .89, GFI= .88, NFI= .88, RMSEA= .15)(see Table 2).
The 2 factor model comprising Emotional dissonance and Perceived Organization Support (χ²
=72.88, df=39, p≤ .001; CFI =.98 , GFI =.96, NFI = .96 , RMSEA = .05) displayed superior
model fit than the one factor model (χ² =324.86, df=34, p<0.001; CFI= .84, GFI= .84, NFI= .83,
RMSEA= .16). Lastly, the 2 factor model of Emotional dissonance and perceptions of co-worker
support (χ² =37.45, df=16, p<0.001; CFI=.98 , GFI =.97, NFI= .97 , RMSEA= .06) also
76
exhibited higher model fit than the one factor model (χ² =56.78, df=16, p<0.001; CFI= .96, GFI=
.96, NFI= .95, RMSEA= .09)(see Table 2).
Time 3 Analysis
For time 3 since mediators including emotional dissonance was tapped at time 3 as well therefore
similar CFA comparisons of emotional dissonance with other mediators were conducted at time
3 as well to establish discriminant validity. A four factor model -factor model comprising
Emotional Dissonance, Perceived Organization Support, supervisor perceptions of support and
perceptions of coworker support at time-3 (χ² =304.82, df=139, p≤ .001; CFI =.96 , GFI =.91 ,
NFI = .93 , RMSEA = .06) demonstrated enhanced model fit than the one factor model (χ²
=1892.65, df=152, p<0.001; CFI= .57, GFI= .55, NFI= .55, RMSEA= .19)(see Table 3).
The 2 factor paired comparison of Emotional dissonance with each of the other mediator
variables measured at time 3 revealed the superiority of each of the 2 factor models over the
single factor model establishing discriminant validity. The 2 factor model comprising Emotional
dissonance and Perceived Supervisor Support at time 3 (χ² =55.24, df= 30, p≤ .001; CFI =.99,
GFI =.97, NFI = .97 , RMSEA= .05) demonstrated higher model fit than the one factor model (χ²
= 1189.75, df=35, p<0.001; CFI= .38, GFI= .53, NFI= .38, RMSEA= .32)(see Table 3).
The 2 factor model comprising Emotional dissonance and Perceived Organization Support (χ²
=86.67, df=38, p≤ .001; CFI =.98 , GFI =.95, NFI = .96 , RMSEA = .06) displayed superior
model fit than the one factor model (χ² =839.81, df=44, p<0.001; CFI= .60, GFI= .62, NFI= .59,
RMSEA= .06). Lastly, the 2 factor model of Emotional dissonance and perceptions of supervisor
77
and coworker support (χ² =47.53, df=16, p<0.001; CFI=.98 , GFI=.97, NFI= .96 , RMSEA= .08)
also demonstrated enhanced model fit than the one factor model (χ² =595.91, df=20, p<0.001;
CFI= .55, GFI= .71, NFI= .55, RMSEA= .29)(see Table 3).
3.8.2 Perceived Support Types
3.8.2.1 Perceived Organizational Support
Time 2 Analysis
A three factor model factor model comprising Perceived Organization Support, perceptions of
supervisory and coworker support at time 2 (χ² =196.50, df=70, p<0.001; CFI=.96 , GFI=.93 ,
NFI= .94 , RMSEA= .07) demonstrated higher model fit than the one factor model (χ² =428.32,
df=61, p<0.001; CFI= .88, GFI= .82, NFI= .86, RMSEA= .13)(see Table 2).
The 2 factor paired comparison of Perceived Organization Support with each of the other
mediator variables measured at time 2 revealed the superiority of each of the 2 factor models
over the single factor model establishing discriminant validity. The 2 factor model comprising
Perceived Organization Support and Perceived Supervisor Support (χ² =102.62, df=39, p≤ .001;
CFI =.97 , GFI =.95, NFI = .96 , RMSEA = .07) demonstrated higher model fit than the one
factor model (χ² = 96.96, df=24, p<0.001; CFI= .97, GFI= .95, NFI= .96, RMSEA= .10)(see
Table 2).
The 2 factor model comprising Perceived Organization Support and Perceived Co-worker
Support (χ² =51.93, df=23, p<0.001; CFI=.98 , GFI=.97, NFI= .97 , RMSEA= .06) displayed
superior model fit than the one factor model (χ² =107.59, df=24, p<0.001; CFI= .95, GFI= .93,
NFI= .93, RMSEA= .10) (see Table 2).
78
Time 3 Analysis
A three factor model factor model comprising Perceived Organization Support, supervisor and
coworker perceptions of support at time 3 (χ² =142.69, df=66, p<0.001; CFI=.98 , GFI=.94 , NFI
= .96 , RMSEA= .06) demonstrated higher model fit than the one factor model (χ² =972.29,
df=77, p<0.001; CFI= .72, GFI= .64, NFI= .71, RMSEA= .19)(see Table 3).
The 2 factor paired comparison of Perceived Organization Support with each of the other
mediator variables measured at time 3 highlighted the superiority of each of the 2 factor models
over the single factor model establishing discriminant validity. The 2 factor model comprising
Perceived Organization Support and Perceived Supervisor Support at time 3 (χ² =89.77, df=39,
p<0.001; CFI=.97 , GFI=.95, NFI= .97 , RMSEA= .06) demonstrated increased model fit than
the one factor model (χ² =580.32, df=44, p<0.001; CFI= .78, GFI= .68, NFI= .77, RMSEA=
.19)(see Table 3).
The 2 factor model comprising Perceived Organization Support and Perceived Co-worker
Support (χ² =46.82, df=23, p≤ .001; CFI =.99 , GFI =.97, NFI = .98 , RMSEA = .06)
demonstrated improved model fit than the one factor model (χ² =348.92, df=20, p<0.001; CFI=
.81, GFI= .79, NFI= .80, RMSEA= .22) (see Table 3).
3.8.2.2 Perceptions of coworker and supervisor support
Time 2 Analysis
The 2 factor model comprising supervisor and coworker perceptions of support at time 2 (χ²
=40.29, df=15, p≤ .001; CFI =.98 , GFI =.97, NFI = .98 , RMSEA = .07) demonstrated improved
79
model fit than the one factor model (χ² =41.40, df=15, p<0.001; CFI= .98, GFI= .97, NFI= .97,
RMSEA= .07) (see Table 2).
Time 3 Analysis
The 2 factor model comprising perceptions of coworker and supervisor support at time 3 (χ²
=52.65, df=16, p≤ .001; CFI =.98 , GFI =.96, NFI = .97 , RMSEA = .08) demonstrated enhanced
model fit than the one factor model (χ² =348.92, df=20, p<0.001; CFI= .81, GFI= .79, NFI= .80,
RMSEA= .22) (see Table 3).
3.8.3 Job Stress
Time 3 Analysis
The outcome variables in the present study were tapped at time 3 only therefore CFA
comparisons of 4 factor, 3 factor and then 2 factor pairing for each of the outcome variables was
done at time 3 to ascertain discriminant validity. A four factor model -factor model comprising
Job stress, burnout, Turnover intensions and Psychological strain at time-3 (χ² =1165.97, df=502,
p≤ .001; CFI =.92 , GFI =.83 , NFI = .87 , RMSEA = .06) demonstrated improved model fit than
the one factor model (χ² =3606.43, df=527, p<0.001; CFI= .62, GFI= .47, NFI= .59, RMSEA=
.13)(see Table 3).
A 3 factor model comprising of Job stress, Burnout and Psychological strain at time 3 (χ²
=1011.68, df= 439, p≤ .001; CFI =.93, GFI =.85, NFI = .88 , RMSEA = .06) demonstrated
80
enhanced model fit than the one factor model (χ² = 3283.39, df=464, p<0.001; CFI= .64, GFI=
.48, NFI= .60, RMSEA= .14)(see Table 3).
The 2 factor model comprising Job Stress and Burnout (χ² =370.77, df=157, p≤ .001; CFI =.95,
GFI =.90, NFI = .92 , RMSEA = .06) displayed superior model fit than the one factor model (χ²
=1611.55, df=170, p<0.001; CFI= .66, GFI= .55, NFI= .64, RMSEA= .16). The 2 factor model
of Job stress and Turnover Intensions (χ² =188.71, df=77, p≤ .001; CFI =.96 , GFI =.93, NFI =
.93, RMSEA = .06) also demonstrated higher model fit than the one factor model (χ² =577.77,
df=90, p<0.001; CFI= .82, GFI= .81, NFI= .79, RMSEA= .13)(see Table 3). The 2 factor model
of Job stress and Psychological Strain (χ² =653.74, df=255, p<0.001; CFI=.93 , GFI=.87, NFI=
.89, RMSEA= .07) also exhibited higher model fit than the one factor model (χ² =2371.72,
df=275, p<0.001; CFI= .63, GFI= .46, NFI= .60, RMSEA= .15)(see Table 3).
3.8.4 Burnout
Time 3
The 2 factor model of Job burnout and Turnover Intensions (χ² =43.94, df=18, p≤ .001; CFI =.99,
GFI =.97, NFI = .98, RMSEA = .07) also demonstrated higher model fit than the one factor
model (χ² =455.02, df=27, p<0.001; CFI= .80, GFI= .78, NFI= .79, RMSEA= .22)(see Table 3).
The 2 factor model of Job burnout and Psychological Strain (χ² =435.18, df=138, p≤ .001; CFI
=.94, GFI =.88, NFI = .92, RMSEA = .08) also exhibited improved model fit than the one factor
model (χ² =1335.40, df=152, p<0.001; CFI= .77, GFI= .65, NFI= .75, RMSEA= .15)(see Table
3).
81
3.8.5 Psychological Strain and Turnover Intensions
Time 3
The 2 factor model of Psychological Strain and Turnover Intensions (χ² =237.27, df=65, p≤ .001;
CFI =.95, GFI =.91, NFI = .93, RMSEA = .09) also higher model fit than the one factor model
(χ² =737.27, df=77, p<0.001; CFI= .80, GFI= .76, NFI= .79, RMSEA= .16)(see Table 3).
Apart from the above possible paired comparison of CFAs tested of variables measured within
the same time period and same source I also ran a full CFA model (measurement model) of all
the study variables (i.e. 9 factor model which is conducted as an initial step in the Structural
Equation Modeling).
Table 2. Results of Confirmatory Factor Analyses for Time-2 Variables
Measurement Models χ² Df CFI GFI NFI RMSEA
ED - POS -PSS - PCS (4 Factor Model)
ED - POS - PSS - PCS (1 Factor Model)
397.74 145 .93 .88 .89 0.07
1830.77 152 .55 .55 .53 0.18
POS -PSS -PCS (3 Factor Model)
POS -PSS -PCS (1 Factor Model)
196.50 70 .96 .93 .94 0.07
428.32 61 .88 .82 .86 0.13
POS -PSS (2 Factor Model)
POS -PSS (1 Factor Model)
102.62 39 .97 .95 .96 0.07
96.96 24 .97 .95 .96 0.10
POS- PCS (2 Factor Model)
POS- PCS (1 Factor Model)
51.93 23 .98 .97 .97 0.06
107.59 24 .95 .93 .93 0.10
PSS- PCS (2 Factor Model) 40.29 15 .98 .97 .98 0.07
82
PSS- PCS (1 Factor Model) 41.40 15 .98 .97 .97 0.07
ED -PSS (2 Factor Model)
ED -PSS (1 Factor Model)
82.89 31 .97 .95 .95 0.07
213.82 26 .89 .89 .88 0.15
ED - POS (2 Factor Model)
ED - POS (1 Factor Model)
72.88 39 .98 .96 .96 0.05
324.86 34 .84 .84 .83 0.16
ED - PCS (2 Factor Model)
ED - PCS (1 Factor Model)
37.45 16 .98 .97 .97 0.06
56.78 16 .96 .96 .95 0.09
Table 3. Results of Confirmatory Factor Analyses for Time-3 Variables
Measurement Models χ² Df CFI GFI NFI RMSEA
ED - POS -PSS - PCS (4 Factor Model)
ED - POS - PSS - PCS (1 Factor Model)
304.82 139 .96 .91 .93 0.06
1892.65 152 .57 .55 .55 0.19
POS -PSS -PCS (3 Factor Model)
POS -PSS -PCS (1 Factor Model)
142.69 66 .98 .94 .96 0.06
972.29 77 .72 .64 .71 0.19
POS -PSS (2 Factor Model)
POS -PSS (1 Factor Model)
89.77 39 .98 .95 .97 0.06
580.32 44 .78 .68 .77 0.19
POS- PCS (2 Factor Model)
POS- PCS (1 Factor Model)
46.82 23 .99 .97 .98 0.06
429.67 27 .78 .78 .77 0.21
PSS- PCS (2 Factor Model)
PSS- PCS (1 Factor Model)
52.65 16 .98 .96 .97 0.08
348.92 20 .81 .79 .80 0.22
ED -PSS (2 Factor Model)
ED -PSS (1 Factor Model)
55.24 30 .99 .97 .97 .05
1189.75 35 .38 .53 .38 .32
83
ED - POS (2 Factor Model)
ED - POS (1 Factor Model)
86.67 38 .98 .95 .96 .06
839.81 44 .60 .62 .59 .23
ED - PCS (2 Factor Model)
ED - PCS (1 Factor Model)
47.53 16 .98 .97 .96 .08
595.91 20 .55 .71 .55 .29
Stress- BO- TOI - Strain (4 Factor Model)
Stress- BO- TOI - Strain (1 Factor Model)
1165.97 502 .92 .83 .87 0.06
3606.43 527 .62 .47 .59 0.13
Stress- BO - Strain (3 Factor Model)
Stress- BO - Strain (1 Factor Model)
1011.68 439 .93 .85 .88 0.06
3283.39 464 .64 .48 .60 0.14
Stress- BO (2 Factor Model)
Stress- BO (1 Factor Model)
370.77 157 .95 .90 .92 0.06
1611.55 170 .66 .55 .64 0.16
Stress - Strain (2 Factor Model)
Stress -Strain (1 Factor Model)
653.74 255 .93 .87 .89 0.07
2371.72 275 .63 .46 .60 0.15
Stress- TOI (2 Factor Model)
Stress- TOI (1 Factor Model)
188.71 77 .96 .93 .93 0.06
577.77 90 .82 .81 .79 0.13
BO - Strain (2 Factor Model)
BO - Strain (1 Factor Model)
435.18 138 .94 .88 .92 0.08
1335.40 152 .77 .65 .75 0.15
BO- TOI (2 Factor Model)
BO- TOI (1 Factor Model)
43.94 18 .99 .97 .98 0.07
455.02 27 .80 .78 .79 0.22
TOI - Strain (2 Factor Model)
TOI - Strain (1 Factor Model)
237.27 65 .95 .91 .93 0.09
737.27 77 .80 .76 .79 0.16
Chapter Summary
84
This chapter provided the reader with the details of the research methods employed in the present
thesis for testing the proposed research model. The chapter provided in-depth information
regarding the research design, target population, sample size planned and attained, sampling
technique used to collect data, data collection procedures used as it was time lagged study,
sample characteristics. The details of how and which of the research model variables were
measured at three time periods i-e time 1, time 2 and time 3 and the appropriate time lag were
provided. Moreover, details regarding the testing of these scales are also given including tests of
reliability and validity. Reliability analysis and test-retest reliability was carried out for variables
measured at two different time periods. The results of confirmatory factor analysis for alternate
models of variables measured from the same source and within one time period were run to
establish discriminant validity which provided superior model fit statistics for the possible
pairing of variables tested in the study.
85
CHAPTER 4
RESULTS
Chapter Overview
This chapter provides the results of data analytical techniques applied in three sections. The first
section gives the results of descriptive statistics and correlation analysis (Pearson bivariate
correlation and paired correlation comparisons) for the model variables. The second section
provides details of Structural Equation Modeling which is done in recommended two major
steps. SEM is conducted by testing Full measurement model (CFA of all the model variables)
and then the confirmation of CFA leads to running of structural paths. The structural paths for
testing direct effects, simple mediation, parallel mediation and sequential mediation are run to
test proposed hypothesis. The last section provides details of proposed hypothesis tested through
the Regression method using Process technique (Preacher & Hayes, 2004) with bootstrapping.
4.1 Descriptive Statistics and Correlation Analysis
Table 4 provides the results of descriptive statistics and co-relation analysis for all the study
variables. All correlations which were above the 0.10 value were significant at p<0.5 and p<0.01
(2 tailed). The means for Workplace Bullying were (M=2.56, SD=1.24), Emotional Dissonance
(M=3.84, SD=1.16), Perceived Organization Support (M=5.00, SD=1.21), Perceived Co-worker
Support (M=5.09, SD=1.39), Perceived Supervisor Support (M=5.19, SD=1.16), Core Self
Evaluations (M=4.68, SD=0.94), Job Stress (M=3.82, SD=1.22), Burnout (M=3.08, SD=1.37),
Psychological Strain (M=2.93, SD=1.36) and Turnover Intensions (M=3.74, SD=1.56).
86
Regarding the correlations between workplace bullying and outcome variables, workplace
bullying had significant positive associations with job stress (r=.33), job burnout (r=.43),
Psychological Strain (r=.54) and Turnover Intensions (r=.21). With respect to the associations
between bullying and mediator variables in the study, Workplace Bullying had significant
positive association with Emotional Dissonance (r=.30). Workplace Bullying had significant
negative correlations with the three support types i.e. with Perceived Organization Support (r=-
.26), Perceived Co-worker Support (r=-.32) and Perceived Supervisor Support (r=-.37).
Concerning the relationship of Workplace bullying with the study's moderator variable,
workplace bullying displayed significant negative correlations with core self evaluations (r=-.33)
(see Table 4).
Regarding the associations of mediator variables with outcomes, emotional dissonance was
positively and significantly associated with job stress (r=.47), job burnout (r=.37), psychological
strain (r=.38) and Turnover Intensions (r=.33). Perceived Organization Support was negatively
and significantly associated with job stress (r= -.17), job burnout (r= -.37), psychological strain
(r= -.30) and Turnover Intensions (r= -.22). Perceptions of supervisor support was negatively
and significantly related to job stress (r= -.11), job burnout (r= -.35), psychological strain (r= -
.35) and Turnover Intensions (r= -.14). Perceived Co-worker Support was negatively and
significantly associated with job stress (r= -.11), job burnout (r= -.26) and psychological strain
(r= -.31), whereas unrelated to turnover intensions (r= -.01) (see Table 4).
With regard to the association of study's moderator variable with mediator variables, Core self
evaluation had significant negative associations with Emotional Dissonance (r= -.25), whereas
Core self evaluations demonstrated significant positive associations with Perceived Organization
Support (r= .47), Perceived Co-worker Support (r= .30) and Perceived Supervisor Support (r=
87
.37). Core Self Evaluations also displayed significant negative correlations with job stress (r= -
.31), job burnout (r= -.46), psychological strain (r= -.49) and Turnover Intensions (r= -.16) (see
Table 4).
Concerning the associations among the mediator variables Emotional Dissonance was
significantly negatively related to Perceived Organization Support (r= -.46) only whereas it was
unrelated to the other two support types i.e. perceptions of supervisor support (r= -.01) and
perceptions of co-worker support (r= .06). Regarding the correlations among the outcome
variables Job Stress was significantly positively related to burnout (r=.57), psychological strain
(r= .52) and turnover intensions (r= .45). Job Burnout demonstrated significant associations with
turnover intensions (r= .46) and psychological strain (r= .75) and Psychological strain had
significant positive associations with Turnover Intensions (r= .33) (see Table 4).
4.1.1 Paired Co-relation Comparisons of Time-2 and Time 3 variables
I also conducted paired correlation comparisons of mediator variables measured at time 2 and
time 3. Table 5 depicts the associations among each pairs of the mediator variables at time 2 and
time 3. With respect to the associations of each similar pairs, Emotional dissonance at time-2 had
significantly positive correlations with Emotional dissonance at time-3 (r= .73). Perceived
Organization Support at time-2 depicted strong positive associations with Perceived
Organization Support at time-3 (r= .74). Similarly, Perceived Supervisor Support at time-2 had
significant positive correlations with Perceived Supervisor Support at time -3 (r= .80). Lastly,
Perceived Co-worker support at time-2 displayed significant positive correlations with Perceived
Co-worker support at time-3 (r= .75).
88
Table 4. Means, Standard deviations, Correlations, and Reliabilities for main variables of interest in the study
Note. N=334; Control variables are age and gender; for Gender 0= ‘Male’ and 1= ‘Female’; WPB=Workplace Bullying;
ED=Emotional Dissonance; POS= Perceived Organizational Support; PSS= Perceived Supervisor Support; PCS= Perceived Co-
worker Support; CSE= Core Self Evaluations, Alpha reliabilities presented in parenthesis. *p<.05, **p<.01, ***p<.001.
Variable Mean SD 1 2 3 4 5 6 7 8 9 10 11 12
1. Age 30.13 8.58
2. Gender .40 .49 -.29
3. WPB (Time-1) 2.56 1.24 .18** .03 (.88)
4. ED (Time-2) 3.84 1.16 .07 .06 .30** (.84)
5. POS (Time-2) 5.00 1.21 -.05 .04 -.26** -.13* (.90)
6. PSS (Time-2) 5.19 1.16 -.05 .00 -.37** -.01 .57** (.89)
7. PCS (Time-2) 5.09 1.39 -.07 -.00 -.32** .06 .44** .65** (.84)
8. CSE (Time-2) 4.68 .94 -.08 -.07 -.33** -.25** .47** .37** .30** (.81)
9. Job Stress (Time-3) 3.82 1.22 .04 .02 .33** .47** -.17** -.11* -.11* -.31** (.92)
10. Burnout (Time-3) 3.08 1.37 .05 .10 .43** .37** -.37** -.35** -.26** -.46** .57** (.92)
11. TOI (Time-3) 3.74
1.56 -.01 .04 .21** .33** -.22** -.14* .01 -.16** .45** .46** (.82)
12. Psychological Strain (Time-3) 2.93
1.36 .15** .05 .54** .38** -.30** -.35** -.31** -.49** .52** .75** .33** (.94)
89
Regarding the correlations of variables at time-2 Emotional dissonance was significantly and
negatively correlated with Perceived Organization Support (r= -.13) whereas was unrelated with
perceived supervisory support (r= -.01) and perceived coworker support (r= .06) at time-2. With
regard to the association of mediator variables at time-2 with each pairs at time-3, Emotional
Dissonance at time-2 was negatively and significantly associated with perceived Organization
support (r= -.14) whereas was unrelated with perceptions of supervisor support (r= .02) and
perceptions of co-worker support (r= .10) at time-3. Concerning the associations of these
variables at time 3 showed that Emotional Dissonance at time-3 was negatively and significantly
associated with perceptions of organization support (r= -.16) whereas was unrelated with
Perceived Supervisor Support (r= -.07) and Perceived Co-worker Support (r= .03) at time-3.
Table 5. Paired Co-relation Comparisons of Variables measured at Time-2 and Time-3
Variables 1 2 3 4 5 6 7 8
1. ED (Time-2) -
2. ED (Time-3) .73** -
3. POS (Time-2) -.13* -.19** -
4. POS (Time-3) -.14** -.16** .74** -
5. PSS (Time-2) -.01 -.06 .57** .58** -
90
6. PSS (Time-3) .02 -.07 .50** .64** .80** -
7. PCS (Time-2) .06 -.02 .44** .44** .65** .53** -
8. PCS (Time-3) .10 .03 .35** .50** .56** .62** .75** -
ED=Emotional Dissonance; PSS=Perceived Supervisory Support; POS=Perceived Organization
Support and PCS= Perceived Co-worker Support
*p<0.05
**p<0.01
***p<0.001
4.2 Structural Equation Modeling Analysis
In the present thesis, I employed Structural Equation Modeling technique using Analysis of
Moment Structures (AMOS) software version 18 to test each of my hypotheses stating direct,
simple mediation, parallel and sequential mediation paths. SEM or Structural Equation Modeling
is a statistical procedure which helps the researcher to test and affirm a sequence of structural
paths i.e. regression equations. The structural models are conceived on the basis of theory/ies and
are then pictorially drawn in the software to help the researcher build up a clearer thought of the
primary casual mechanisms under investigation (Byrne, 2013). The hypothesized structural paths
can then be statistically verified through examination of the one to one paths and at the same
time of the complete variables in the model to confirm the extent to which these structural
models adequately matches the data. Based on the goodness of fit statistics it is established if the
91
hypothesized relationships among the model variables are confirmed or disapproved. In addition,
the label of parameter estimates section depicts the beta values or standardized regression
weights of the individual structural paths along with their level of significance (p values) which
provides additional credibility for the substantiation of hypothesis.
SEM is considered to be extremely dependable and is more advanced than other multivariate
approaches to data analysis. To begin with, SEM adopts a confirmatory rather than exploratory
approach for performing data analysis. Another distinguishing feature of Structural equation
modeling is that it permits to check a series of structural paths separately and concurrently thus
making hypothesis testing simpler yet complete. Further, the calculation of individual parameter
estimates (beta values) of paths and complete range of model fit statistics values enhances the
worth of this approach. Lastly, by employing the SEM technique, full measurement models or
CFAs can be tested and verified which reduces errors thus making hypothesized relations among
unobserved or latent variables less influenced by the presence of measurement errors.
The fundamental Structural equation modeling is bifurcated into two parts e.g. full measurement
model and the structural model. Hence, the process which I pursued was a two stage procedure in
which first a full CFA model (measurement model) was run and substantiated. The confirmation
of full CFA model led towards analyzing and running of the structural paths so as to verify the
hypothesis (Fitzgerald, Drasgow, Hulin, Gelfand & Magley, 1997; Little, Cunningham, Shahar
& Widaman, 2002; Kline, 2005).
92
The full CFA model also called measurement model is defined as a model which includes the
relationships among the unobserved or latent variables alongside the manifest or observed
variables. Therefore, a full CFA model determines the relationship between scale items of a
measured instrument called the manifest or observed variables along with the proposed
constructs which they are tapping called as latent or unobserved variables. Following the
substantiation of the measurement model, then the structural models are run to confirm
hypothesis. The structural models also called as the path models determines the association
between the latent variables. The structural path models highlights the manner in which
particular unobserved variables alongside their manifest or observed variables either directly or
indirectly affects or brings about alterations in the values of other unobserved variables in the
model. In other words, the structural model sequentially illustrates the paths as characterized by
the independent also called exogenous variables causing change in the intervening and dependant
variables labeled as endogenous variables (Byrne, 2013).
4.2.1 Full Measurement Model (Full CFA model)
Confirmatory factor analysis (CFA) was performed to check and authenticate the full
measurement model consisting of the 9 constructs (i-e Workplace Bullying, Emotional
dissonance, perceived co-worker support, perceived organization support, perceived supervisory
support, Job stress, Psychological strain, Burnout and turnover intensions) under investigation.
The Full CFA model consisting of all the present thesis variables was run to ascertain the
discriminant and convergent validity of the measures. The full measurement model fit with the
93
data was assessed through the goodness of fit, chi-square and the model fit statistics. The model
goodness of fit i.e. chi-square should be significant (p<0.05) which signals adequate fit.
The Fit indexes were checked through the Confirmatory fit index (CFI), Normed Fit index (NFI)
and Goodness of fit index (GFI) values greater than 0.9 signifies adequate model fit. Then the
root mean square error of appropriation (RMSEA) and root mean square residual value (SRMR)
is also used to authenticate appropriate model fit and a value less than 0.08 is considered
acceptable and indicates appropriate model fit (Bentler & Bonett, 1980; Bollen, 1989; Joreskog
& Sorbom, 1999). The model fit statistics showed that the measurement model (Full CFA model)
adequate fits the data χ² = 3262.89, df=1701, p<0.001; CFI= 0.90 , GFI= 0.78, NFI= 0.80,
RMSEA= 0.05 demonstrating suitable model fit and hence supporting the full factor analytical
model.
In order to test convergent validity of the measures, the average variance extracted (AVE) and
the composite reliabilities were calculated. All the 9 constructs had AVE above the minimum 0.5
threshold. The composite reliability (CR) scores for all the 9 constructs were above the
acceptable values of 0.6 (Fornell & Larcker, 1981). The AVE values of perceptions of
organization support = 0.58, perceptions of co-worker support= 0.60, perceptions of supervisory
support= 0.62, Workplace Bullying= 0.50, Emotional Dissonance= 0.49, Psychological Strain=
0.59, Stress= 0.50, Burnout=0.63, Turnover Intensions= 0.75 indicate that they fall under the
acceptable limit. Moreover, the results of the composite reliability scores Perceived
Organizational Support= 0.89, Perceived Supervisor Support= 0.89, Perceived Co-worker
support= 0.81, Workplace Bullying= 0.87, Emotional Dissonance= 0.83, Psychological Strain=
94
0.95, Stress= 0.92, Burnout=0.92, Turnover Intensions= 0.85 demonstrated excellent fit
indicating convergent validity for all these 9 constructs. The findings of the full CFA model also
demonstrated factor loadings greater than 0.5 and more (refer to Fig 2 for factor loadings of each
item of the construct) (Anderson & Gerbing, 1988).
Figure 2 depicts the full measurement model comprising the factor loadings of each item drawn
to portray the full CFA and below this figure the full measurement model as given in the AMOS
output file. (This model was drawn since the values were not clear in the AMOS path diagram
builder).
95
Workplace
Bullying
Emotional
Dissonance
Perceived
Organization
Support
Perceived
Co-worker
Support
Perceived
Supervisor
Support
Psychological
Strain
Job Stress
Turnover
Intensions
Burnout
WB1T1
WB2T1
WB3T1
WB4T1
WB5T1
WB6T1
WB7T1
WB8T1
e1
e2
e3
e4
e5
e6
e7
e8
POS2T2 POS1T2 POS8T2 POS5T2 POS4T2 POS3T2
ED5T2
ED4T2
ED3T2
ED2T2
ED1T2
e16
e15
e17
e18
e19
e9
e10
e11
e12
e13
e14
PStrain1T3
PStrain2T3
PStrain3T3
PStrain4T3
PStrain5T3
PStrain6T3
PStrain7T3
PStrain8T3
PStrain9T3
PStrain10T3
PStrain11T3
PCS2T2
PCS3T2
PCS1T2
JS1T3
JS13T3
JS12T3
JS11T3
JS2T3
JS3T3
JS4T3
JS5T3
JS6T3
JS7T3
JS8T3
JS9T3
JS10T3
PStrain12T3
TO1T3
TO2T3
e58
e59
e57
e61
e60
BO2T3
BO1T3
BO3T3
BO5T3
BO4T3
e20
e21
e22
e23
e24
e25
e26
e27
e28
e29
e30
e31
BO6T3
BO7T3
PSS4T3 PSS7T3
PSS1T3 PSS2T3 PSS3T3
e33
e32
e36
e35
e34
e42
e41
e40
e39
e38
e37
e43
e44
e45
e46
e47
e48
e49
e50
e51
e52
e53
e54
e55
e56
.41
.51
.60
.67
.81
.86
.74
.76
.84
.86 .84 .77
.54
.68
.70
.74
.73
.68
.66
.91 .79 .60
.70 .98
.85
.86 .78
.79 .65
.68
.80
.81
.85
.85
.82
.71
.58
.77
.75
.81
.80
.84
.76 .79
.84
.76 .75
.74
.61
.72
.71
.81
.76
.70
.75
.73
.74
.46
.55
.67
.68
96
Figure 2. Full Measurement Model (AMOS Path Diagram)
97
4.2.2 Direct Structural Model
4.2.2.1 Path model of Workplace Bullying relationship to Job Stress, Burnout, Psychological
Strain and Turnover Intensions
The analysis and successive confirmation of the full CFA model (measurement model) then
resulted in the second stage whereby path models for checking direct effects, simple, parallel,
sequential mediation were run for verifying hypothesis. The structural model for testing the main
effects was run for Hypothesis 1 which postulated that workplace bullying is positively
associated with Job Stress, Burnout, Psychological Strain and Turnover Intensions.
The structural model diagram for H1 is demonstrated in Fig 3. The model fit statistics for
structural model testing the direct effects of workplace bullying on outcomes i.e. Chi-square, χ²
= 1705.54; Df=789; Chi-square/DF ratio= 2.16; p<0.001; CFI= 0.91; and GFI= 0.81; and NFI=
0.84; and RMSEA= 0.06 demonstrated that the direct effects model adequately fits the data.
Moreover, the results show that the structural paths from Workplace bullying to Job Stress
(β=0.47, p<0.001), Turnover Intensions (β=0.41, p<0.001), Psychological Strain (β=0.69,
p<0.001) and Burnout (β=0.66, p<0.001) were significant and positive. Workplace bullying
explained 22% variation in job stress, 17% in turnover intensions, 48% in psychological strain
and 43% in Burnout. Thus H1 was fully supported.
98
Figure 3. Path model of Workplace Bullying related to Stress, Burnout, Turnover
Intensions and Psychological Strain
99
4.2.2.2 Path model of Perceived Organization Support relationship to Job Stress, Burnout,
Psychological Strain and Turnover Intensions
Hypothesis 2a predicted that Perceived Organization support is negatively associated with job
Stress, Burnout, Psychological Strain and Turnover Intensions.
The structural model diagram for H2a is depicted in Figure 4. The model fit statistics for
structural model testing the direct effects of Perceived Organization Support on Job strains i.e.
Chi-square, χ² =1619.45 ; df= 712; chi-square/Df ratio= 2.27; p< .001; CFI= 0.90; and GFI=
0.81; and NFI= 0.84; and RMSEA= 0.06 shows that the direct effects model adequately fits the
data.
The parameter estimate values comprising of individual paths standardized regression weights
(Beta values) and their significance (p values) further corroborate the adequacy of this
hypothesized structural model. The results show that the structural paths from Perceived
Organization Support to Job Stress (β=-0.19, p<0.05), Turnover Intensions (β=-0.28, p<0.001),
Psychological Strain (β=-0.35, p<0.001) and Burnout (β=-0.43, p<0.001) was negative and
significant. Perceived Organization Support explained 3.5% variation in job stress, 7.6% in
turnover intensions, 12% in psychological strain and 18.4% in Burnout. Thus H2a was fully
confirmed.
100
Figure 4. Path model showing Perceived Organization Support's direct relationship to
Stress, Burnout, Turnover Intensions and Psychological Strain
101
4.2.2.3 Path model of Perceived Co-worker Support relationship to Job Stress, Burnout,
Psychological Strain and Turnover Intensions
Hypothesis 2b predicted that perceived coworker support is negatively associated to Job Stress,
Burnout, Psychological Strain and Turnover Intensions.
The structural model fit diagram for H2b is shown in Figure 5. The model fit statistics for
structural model testing the direct paths of perceptions of co-worker support on Job strains i.e.
Chi-square, χ² =130.68 ; df=593 ; chi-square/Df ratio=2.20 ; p< .001; CFI= 0.92; and GFI=
0.83; NFI= 0.86; & RMSEA= 0.06 shows that the direct effects model adequately fits the data.
The values of parameter estimates comprising of standardized regression co-efficients, and
consequent significance (p values) partially support the suitability of this hypothesized structural
model. The results show that the structural paths from Perceived Co-worker Support to Job
Stress (β=0.65, p<0.001), Turnover Intensions (β=0.51, p<0.001), Psychological Strain (β=0.85,
p<0.001) and Burnout (β=-0.43, p<0.001) was although significant but was positive. Perceived
Co-worker Support explained 43% variation in job stress, 26% in turnover intensions, 72% in
psychological strain and 88% in Burnout. Thus H2b was partially supported.
102
Figure 5. Path model showing Perceived co-worker Support's direct relationship to Stress,
Burnout, Turnover Intensions and Psychological Strain
103
4.2.2.4 Path model of Perceived Supervisor Support relationship to Job Stress, Burnout,
Psychological Strain and Turnover Intensions
Hypothesis 2c predicted that Perceived Supervisor support is negatively associated with Job
Stress, Burnout, Psychological Strain and Turnover Intensions.
The structural model fit diagram for H2c is shown in Figure 6. The model fit statistics for
structural model testing the direct effects of Perceived Supervisor Support on Job strains i.e Chi-
square, χ² = 1477.27; df=668; chi-square/Df ratio= 2.21; p< .001; CFI= 0.91; and GFI=0.82;
and NFI= 0.85; and RMSEA=0.06 depicts that the direct effects model adequately fits the data.
The values of the parameter estimates comprising of standardized regression weights and
significance values also substantiate the adequacy of this structural model. The results show that
the structural paths from Perceived supervisor Support to Job Stress (β=-0.12, p<0.001),
Turnover Intensions (β=-0.42, p<0.001), Psychological Strain (β=-0.38, p<0.001) and Burnout
(β=-0.42, p<0.001) was significant. Perceived Supervisor Support explained 1.5% variation in
job stress, 5.4% in turnover intensions, 14% in psychological strain and 17% in Burnout. Thus
H2c was supported.
104
Figure 6. Path model showing Perceived Supervisor Support's direct relationship to Stress,
Burnout, Turnover Intensions and Psychological Strain
105
4.2.2.5 Path model of Emotional Dissonance influencing Job Stress, Burnout, Psychological
Strain and Turnover Intensions
Hypothesis 3 predicted that emotional dissonance is positively associated with Job Stress,
Burnout, Psychological Strain and Turnover Intensions.
The structural model fit diagram for H3 is shown in Figure 7. The model fit statistics for
structural model testing the direct effects of Emotional Dissonance on Job strains i.e. Chi-square,
χ² = 1532.96; df=660; chi-square/Df ratio= 2.32; p< .001; CFI= 0.91; and GFI= 0.81; NFI=
0.85; and RMSEA= 0.06 demonstrates that the direct effects model adequately fits the data.
The values of the parameter estimates comprising of standardized regression weights and
significance values also substantiate the adequacy of this structural model. The results show that
the structural paths from Emotional Dissonance to Job Stress (β=0.67, p<0.001), Turnover
Intensions (β=0.51, p<0.001), Psychological Strain (β=0.61, p<0.001) and Burnout (β=0.54,
p<0.001) was significant. Emotional Dissonance explained 45% variation in job stress, 26% in
turnover intensions, 38% in psychological strain and 29% in Burnout. Thus H3 was fully
supported.
106
Figure 7. Path model showing Emotional Dissonance's direct relationship to Stress,
Burnout, Turnover Intensions and Psychological Strain
107
4.2.2.6 Path model of Workplace Bullying influencing Perceived Organization Support,
Perceived Co-worker Support and Perceived Supervisor Support
Hypothesis 4 (a, b and c) predicted that Workplace Bullying is related negatively to perceived
organization support, perceived Co-worker Support and perceived Supervisor Support
respectively.
The structural model fit diagram for H4a, b and c is shown in Figure 8. The model fit statistics
for structural model testing the direct effects of Workplace Bullying on Perceived Support types
i.e. Chi-square, χ² = 622.03; df=193; chi-square/Df ratio= 3.22; p< .001; CFI= 0.90; and GFI=
0.86; NFI= 0.86; & RMSEA= 0.08 shows that the direct effects model adequately fits the data.
The values of the parameter estimates comprising of standardized regression weights and
significance values also substantiate the adequacy of this structural model. The results show that
the structural paths from Workplace Bullying to Perceived Organization Support (β=-0.15,
p<0.05) & Perceived Co-worker support (β=-0.13, p<0.05) was negative and significant. The
structural path from Workplace Bullying to Perceived Supervisor Support was positive and
insignificant (β=0.05, p= n.s.). Workplace Bullying explained 2.2% variation in Perceived
Organization Support and 1.8% in Perceived Co-worker Support. However, workplace bullying
did not show any variation in Perceived Supervisor Support. Thus H4 a and b was fully
supported, whereas H4c was rejected.
108
Figure 8. Path model of Workplace Bullying directly related to Perceived Organization
Support, Perceived Co-worker Support and Perceived Supervisor Support
109
4.2.2.7 Path model of Workplace Bullying to Emotional Dissonance
Hypothesis 5 predicted that Workplace Bullying is significantly and positively related to
emotional dissonance.
The structural model fit diagram for H5 is shown in Figure 9. The model fit statistics for
structural model testing the direct effects of Workplace Bullying on Emotional Dissonance i.e.
Chi-square, χ² = 108.76; df=54; chi-square/Df ratio= 2.14; p< .001; CFI= 0.97; and GFI= 0.95;
NFI= 0.95; & RMSEA=0.05 shows that the direct effects model adequately fits the data.
The values of the parameter estimates comprising of standardized regression weights and
significance values also substantiate the adequacy of this structural model. The results show that
the structural paths from Workplace Bullying to Emotional Dissonance (β=0.33, p<0.001) was
positive and significant. Workplace Bullying explained 10.8% variation in Emotional
Dissonance. Thus H5 was fully supported.
110
Figure 9. Path model of Workplace Bullying to Emotional Dissonance
111
4.2.3 Structural Models for Simple Mediation
4.2.3.1 Path model of Perceived Organization Support as Mediator between Workplace Bullying
and Job Strains (Stress, Burnout, Psychological Strain and Turnover Intensions)
Hypothesis 6a predicted that perceived organization support acts as a mediator in the relationship
between Workplace Bullying and Job strains (Stress, Burnout, Psychological Strain and
Turnover Intensions). The structural model fit diagram for H6a is shown in Figure 10. The
model fit statistics for structural model testing the mediating effects of Perceived Organization
Support between Workplace Bullying and Job strains i.e. Chi-square, χ² = 1970.16; df=1039;
chi-square/Df ratio= 1.90; p< .001; CFI= 0.92; and GFI= 0.81; NFI= 0.84; & RMSEA= 0.05
shows that the mediating path model adequately fits the data.
The values of the parameter estimates comprising of standardized regression weights and
significance values also substantiate the adequacy of this structural model. The results show that
the structural paths from Workplace Bullying to Perceived Organization Support (β=-0.34,
p<0.001) was negative and significant. Furthermore, the structural paths from Perceived
Organization Support to Job Stress was insignificant (β=-0.19, p= n.s), whereas for Turnover
Intensions (β=-0.31, p<0.001), Psychological Strain (β=-0.36, p<0.001) and Burnout (β=-0.43,
p<0.001) was significant and negatively associated. Workplace Bullying explained 12%
variation in Perceived Organization Support. Furthermore, Perceived Organization Support
explained 10% variation in turnover intensions, 13% in psychological strain and 19% in
Burnout. Thus the mediation hypothesis for Perceived Organization Support (H6a) between
Workplace Bullying and Job Strains was supported for all outcomes except Job Stress.
112
Figure 10. Path model of Perceived Organization Support as Mediator between Workplace
Bullying and Job Strains (Stress, Burnout, Psychological Strain and Turnover Intensions)
113
4.2.3.2 Path model of Perceived Co-worker Support as Mediator between Workplace Bullying
and Job Strains (Stress, Burnout, Psychological Strain and Turnover Intensions)
Hypothesis 6b predicted that Perceived Co-worker Support mediates the relationship between
Workplace Bullying and Job strains (Stress, Burnout, Psychological Strain and Turnover
Intensions).
The structural model fit diagram for H6b is shown in Figure 11. The model fit statistics for
structural model testing the mediating effects of Perceived Co-worker Support between
Workplace Bullying and Job strains i.e. Chi-square, χ² = 1857.03; df=911; chi-square/Df ratio=
2.04; p< .001; CFI= 0.91; and GFI= 0.80; NFI= 0.84; & RMSEA= 0.06 depicts that the
mediating path model adequately fits the data.
The values of the parameter estimates comprising of standardized regression weights and
significance values also substantiate the adequacy of this structural model. The results show that
the structural paths from Workplace Bullying to Perceived Co-worker Support (β=0.60, p<0.001)
was significant. Furthermore, the structural paths from Perceived Co-worker Support to Job
Stress (β=0.62, p=n.s.) and Turnover Intensions (β=0.48, p=n.s.) was insignificant whereas for
Psychological Strain (β=0.89, p<0.001) and Burnout (β=-0.89, p<0.001) was significant.
Workplace Bullying explained 36% variation in Perceived Co-worker Support. Furthermore,
Perceived Co-worker Support explained 23% variation in turnover intensions, 78% in
psychological strain and 80% in Burnout. Thus the mediation hypothesis of Perceived Co-
worker Support (H6b) between Workplace Bullying and Job Strains was supported for all
outcomes except Job Stress and turnover intensions.
114
Figure 11. Path model of Perceived Co-worker Support as Mediator between Workplace
Bullying and Job Strains (Stress, Burnout, Psychological Strain and Turnover Intensions
115
4.2.3.3 Path model of Perceived Supervisor Support as Mediator between Workplace Bullying
and Job Strains (Stress, Burnout, Psychological Strain and Turnover Intensions)
Hypothesis 6c predicted that perceived supervisory support mediates the relationship between
Workplace Bullying and Job strains (Stress, Burnout, Psychological Strain and Turnover
Intensions). The structural model fit diagram for H6c is shown in Figure 12. The model fit
statistics for structural model testing the mediating effects of Perceived supervisor Support
between Workplace Bullying and Job strains i.e. Chi-square, χ² = 1998.80; df=993; chi-
square/Df ratio= 2.01; p< .001; CFI= 0.91; and GFI= 0.80; NFI= 0.84; and RMSEA= 0.06
depicts that the mediating path model adequately fits the data.
The values of the parameter estimates comprising of standardized regression weights and
significance values also verify the suitability of this hypothesized structural model. The results
show that the structural paths from Workplace Bullying to Perceived Supervisor Support
(β=0.44, p<0.001) was significant. Furthermore, the structural paths from Perceived Supervisor
Support to Job Stress (β=0.12, p= n.s) and Turnover Intensions (β=0.24, p=n.s.), Psychological
Strain (β=0.41, p<0.001) and Burnout (β=-0.43, p<0.001) was significant. Workplace Bullying
explained 19% variation in Perceived Supervisor Support. Furthermore, Perceived Supervisor
Support explained 6% variation in turnover intensions, 17% in psychological strain and 18% in
Burnout. Thus the mediation hypothesis of Perceived Supervisor Support (H6c) between
Workplace Bullying and Job Strains was supported for all outcomes except Job Stress and
turnover intensions.
116
Figure 12. Path model of Perceived Supervisor Support as Mediator between Workplace
Bullying and Job Strains (Stress, Burnout, Psychological Strain and Turnover Intensions)
117
4.2.3.4 Path model of Emotional Dissonance as Mediator between Workplace Bullying and Job
Strains (Stress, Burnout, Psychological Strain and Turnover Intensions)
Hypothesis 7 hypothesized that emotional dissonance as a mediator in the relationship between
Workplace Bullying and Job strains (Stress, Burnout, Psychological Strain and Turnover
Intensions).
The structural model fit diagram for H7 is shown in Figure 13. The model fit statistics for
structural model testing the mediating effects of Emotional Dissonance between Workplace
Bullying and Job strains i.e. Chi-square, χ² = 2052.26; df=987; chi-square/Df ratio= 2.07; p<
.001; CFI= 0.90; and GFI= 0.80; and NFI= 0.83; and RMSEA= 0.06 demonstrates that the
mediating path model adequately fits the data.
The values of the parameter estimates comprising of standardized regression weights and
significance values also confirm the acceptability of this structural model. The results show that
the structural paths from Workplace Bullying to Emotional Dissonance (β=0.54, p<0.001) was
significant. Furthermore, the structural paths from Emotional Dissonance to Job Stress (β=0.68,
p<0.001), Turnover Intensions (β=0.50, p<0.001), Psychological Strain (β=0.74, p<0.001) and
Burnout (β=-0.65, p<0.001) was significant. Workplace Bullying explained 30% variation in
Emotional Dissonance. Furthermore, Emotional Dissonance explained 46% variation in Job
Stress, 25% in Turnover Intensions, 54% psychological strain and 42% in Burnout. Thus the
mediation hypothesis of Emotional Dissonance (H7) between Workplace Bullying and Job
Strains was fully supported for all outcomes.
118
Figure 13. Path model of Emotional Dissonance as Mediator between Workplace Bullying
and Job Strains (Stress, Burnout, Psychological Strain and Turnover Intensions)
119
4.2.4 Path model of Perceived Support Types and Emotional Dissonance as sequential
Mediators between Workplace Bullying and Job Strains (Stress, Burnout, Psychological
Strain and Turnover Intensions)
Hypothesis 8 predicted that Perceived Support types (Perceived Organization, Perceived
supervisor and Perceived coworker support) and Emotional Dissonance act as sequentially
mediators in the relationship between Bullying and Job strains (Stress, Burnout, Psychological
Strain and Turnover Intensions).
The structural model fit diagram for H8 is shown in Figure 14. The model fit statistics for
structural model testing the mediating effects of Emotional Dissonance between Workplace
Bullying and Job strains i.e. Chi-square, χ² = 3593.06; df=1715; chi-square/Df ratio= 2.10;
p<0.001; CFI=0.87; and GFI= 0.74; NFI= 0.78; and RMSEA= 0.06 shows that the sequential
mediation path model adequately fits the data.
The values of the parameter estimates comprising of standardized regression weights and
significance values also verify the suitability of this structural model. The results show that the
structural paths from Workplace Bullying to Perceived Supervisor Support (β=-0.40, p<0.001),
Perceived Co-worker Support (β=-0.36, p<0.001) and Perceived Organization Support (β=-0.33,
p<0.001) was negative and significant. Furthermore, the structural paths from Perceived
Organization Support to Emotional Dissonance (β=-0.23, p<0.001) was significant and negative.
The structural paths of Perceived Supervisor Support to Emotional Dissonance (β=-0.10, p=n.s.),
and Perceived Co-worker Support to Emotional Dissonance (β=0.10, p=n.s.) were insignificant.
The structural paths from Emotional Dissonance to Job Stress (β=0.40, p<0.001), Turnover
Intensions (β=0.48, p<0.001), Psychological Strain (β=0.60, p<0.001) and Burnout (β=0.58,
120
p<0.001) was significant. Workplace Bullying explained 13% variation in Perceived Co-worker
Support, 11% in Perceived Organization Support and 16% in Perceived Supervisor Support.
Furthermore, Emotional Dissonance explained 40% variation in Job Stress, 23% in Turnover
Intensions, 36% psychological strain and 33% in Burnout. Since the structural paths of Perceived
Organization support was significant to Emotional dissonance whereas the other two support
types i.e. Perceived Co-worker and Perceived Supervisor support had insignificant paths with
emotional dissonance, therefore the sequential mediation hypothesis of Perceived organization
support and Emotional Dissonance (H8) between Workplace Bullying and Job Strains was fully
supported for all outcomes. Whereas the sequential mediation of PSS and PCS was not supported
lending partial support for hypothesis 8.
121
Figure 14. Path model of Perceived Support Types and Emotional Dissonance as sequential
Mediators between Workplace Bullying and Job Strains (Stress, Burnout, Psychological
Strain and Turnover Intensions)
122
Figure 15. Full Structural Model of Perceived Support Types (POS, PSS and PCS) and
Emotional Dissonance as Parallel and sequential Mediators between Workplace Bullying
and Job Strains (Stress, Burnout, Psychological Strain and Turnover Intensions)
N=334; Full structural model showing direct effects, simple, parallel and sequential mediating
effects. The actual diagram as displayed in AMOS path diagram builder is depicted below.
Standardized regression weight values on paths and asterisks indicate significance values. Model
Fit: χ² = 3179.52; df=1692; χ²/Df ratio= 1.88; p< .001; CFI= 0.90; and GFI= 0.78; NFI= 0.80;
& RMSEA= 0.05.
(a) Parallel mediation paths running from bullying to emotional dissonance to outcomes on one
hand and bullying to support types to outcomes (Stress, BO, Strain and TOI) on the other.
(b) Sequential mediation paths are reflected from workplace bullying to the 3 support types
(POS, PSS and PCS) to Emotional dissonance to outcomes (Stress, BO, Strain and TOI).
Workplace
Bullying
Emotional
Dissonance
Perceived
Organization
Support
Perceived
Co-worker
Support
Perceived
Supervisor
Support
Job Stress
Psychological
Strain
Job Burnout
Turnover
Intensions
.42***
-.32***
-.40***
-.37***
.48***
.29***
.31*** .34**
*
-.07
-.06
-.21*** -.24***
.10
-.11***
-.15***
-..11 -.14***
-.20*
-.01
-.15***
-.17*
.08
.24
123
Figure 15: Path model of Perceived Support Types (POS, PSS and PCS) and Emotional Dissonance as
Parallel and sequential Mediators between Workplace Bullying and Job Strains (Stress, BO, Strain and TOI)
(AMOS Output)
124
4.3 Mediation and Moderation analysis using Bootstrapping
In the current study, I also conducted regression analysis through the process technique
developed by Preacher & Hayes (2004; 2008). The Process approach is an updated and advanced
data analysis technique which utilizes bootstrapping for conducting and checking moderation
and mediation hypothesis. The results of the structural equation modeling provided support for
majority of the direct, simple mediation, parallel mediation and sequential mediation hypothesis.
Therefore after implementing SEM, I utilized another robust technique of Process to verify my
hypotheses. I first tested simple mediation hypotheses for my 4 mediators i.e Emotional
Dissonance, Perceived Organization Support, Perceived Co-worker Support and Perceived
Supervisor Support. Then tested for parallel and sequential mediation hypothesis. Lastly
conducted moderation analysis using bootstrapping.
4.3.1 Tests of Mediation (Simple, Parallel and Sequential)
According to Baron and Kenny (1986), mediational analysis was conducted on the basis of
fulfilling three pre-requisite conditions. However, the step-wise technique of performing
mediation was marred with numerous weaknesses (Mackinnon, Lockwood, Hoffman, West &
Sheets, 2002).
For example, Baron & Kenny (1986) suggest that three fundamental conditions should be met
before performing mediational analysis. First of all, the main effects of Independent variable (X)
on the dependant variable (Y) should be significant. Next, the relationship between Independent
variable (X) and Mediator (M) should be significant and finally, the mediator variable should be
significantly associated to the outcome or dependant variable. If any of these three pre-requisites
are not fulfilled mediational analysis is not possible. Hence upon sufficiently meeting these
125
requirements, the researcher is able to conduct mediational analysis whereby it is assumed that in
order to confirm full mediation, when the mediator variable is added in the data analysis process
the relationship of independent variable on dependant variable becomes insignificant. This would
imply that the influence of independent variable is caused on the outcome or dependent variable
through the several causal mechanisms of intervening or mediating variable(s) thereby lessening
the main effects of the IV(predictor) on the DV (outcome) variables (Shrout & Bolger, 2002).
Nonetheless, the statistical and research methods experts have been skeptical and raised many
concerns regarding the importance given by Baron and Kenny (1986) to ascertain the
relationship between the IV and DV (Mackinnon, Krull & Lockwood, 2000; Shrout & Bolger,
2002). Responding to these objections, Kenny, Kashy and Bolger (1998) provided latest
suggestions and recognized that the first condition is not essential for performing mediational
analysis. Thus, it was recommended that mediational analysis can be substantiated clearly by
looking into the significance values of the indirect effects which are represented through the
Sobel (1982) test.
In line with this Preacher and Hayes (2004) contended that the above approach is more reliable
as compared to Baron and Kenny's (1986) sequence wise method of mediation, since it more
succinctly illustrates mediation. In addition, it is also argued that the Sobel (1982) test is also not
free of weaknesses as it assumes that there is normal distribution between the indirect effects that
range from the independent variable to the dependent variables. This supposition is flawed as the
as the relationship between X and Y is not normally distributed even when the factors producing
the XY terms are normally dispersed (Edwards & Lambert, 2007). To address this concern
experts have recommended using the bootstrapping technique. By employing confidence
126
intervals which are bootstrapped, it is possible to avoid strength issues generated as a result of
uneven sample distributions and other irregularities of indirect effects (MacKinnon, Lockwood
& Williams, 2004).
Therefore in the present thesis, I utilized the mediational analysis for testing direct, simple
mediation, parallel and sequential mediation as recommended by Preacher and Hayes (2004).
Preacher & Hayes (2004) formulated a macro tool which can be installed in the SPSS software
that guides the researcher in the analysis and calculation of determining indirect effects, by
utilizing Sobel (1982) test. This technique also utilizes bootstrapping method to determine
confidence intervals (CI) and also incorporates the stepwise approach devised by Baron and
Kenny (1986). Preacher and Hayes also provided templates of 74 models which can be used by
the researcher to specify the kind of analysis required and then input study variables in
accordance with the model specified. For testing simple mediation I ran model 4 for the 4
mediators (Emotional Dissonance, Perceived Organization Support, perceived supervisory
support and perceived coworker support) for each of the four outcomes. I then used model 6 for
testing both parallel and sequential mediation hypothesis simultaneously. Lastly I utilized model
1 for testing my moderation hypothesis.
4.3.2 Regression Results for Simple Mediation using Bootstrapping
4.3.2.1 Emotional Dissonance as mediator between Workplace Bullying and Outcomes (Job
Stress, Burnout, Turnover Intensions and Psychological Strain)
Table 6 presents the results of Emotional Dissonance as a mediator between workplace bullying
and outcomes (Job Stress, Burnout, Turnover Intensions and Psychological Strain) and
127
consequently address H1, H3, H5 and H7. According to Table 6 (a, b, c and d) supporting
Hypothesis1 workplace bullying is significantly and positively associated with job stress (B=
0.21, t= 4.19, p < .001), job burnout (B= 0.39, t = 6.88, p < .001), Turnover Intensions (B= 0.17,
t= 2.51, p < .001) and psychological strain (B=0.49, t = 9.51, p < .001). Therefore H1 was fully
supported. In accordance with Table 6 (a, b, c and d) and in line with hypothesis 3 Emotional
dissonance was found to be significantly and positively associated with job stress (B=0.43, t =
8.20, p<0.001), job burnout (B= 0.31, t = 5.22, p < 0.001), Turnover Intensions (B= 0.39, t =
5.41, p < 0.001) and psychological strain (B = 0.28, t = 5.13, p < .001). Thus, Hypothesis 3 was
fully supported. Referring to Table 6 (a, b, c and d) and supporting Hypothesis 5 workplace
bullying was related positively with Emotional dissonance (B= 0.27, t = 5.41, p < 0.001). Thus,
Hypothesis 5 was fully confirmed.
Lastly we tested the mediation hypothesis (Hypothesis 7) for each outcome, which asserts that
Emotional Dissonance acts as a mediator in the relationship between workplace bullying and job
stress, which is supported (B= 0.33, t = 6.22, p < 0.001). The significance tests (two-tailed and
with a normal dispersion) confirmed that the indirect effects were significant (Sobel, z=4.49,
p<0.001). The results of the bootstrap further verified the Sobel test (refer to Table 6a), with
confidence interval (CI) of 95% and the non-zero indirect effects (.07, .18).
Emotional dissonance mediates the relationship between workplace bullying and job burnout
(see Table 6b) which is also supported (B = 0.47, t = 8.40, p < .001). The significance tests (two-
tailed and with a normal dispersion) confirmed that the indirect effects were significant (Sobel,
128
z=3.72, p<0.001). The results of the bootstrap further verified the Sobel test (refer to Table 6b),
with confidence interval (CI) of 95% and the non-zero indirect effects (.04, .14).
Emotional Dissonance mediates the relationship between workplace bullying and Turnover
Intensions (see Table 6c) which is confirmed(B= 0.28, t = 4.06, p < .001). The significance tests
(two-tailed and with a normal dispersion) confirmed that the indirect effects were significant
(Sobel, z=3.79, p<0.001). The results of the bootstrap further verified the Sobel test (refer to
Table 6c), with confidence interval (CI) of 95% and the non-zero indirect effects (.06, .17).
Emotional Dissonance mediates the relationship between workplace bullying and Psychological
Strain (see Table 6d) which is supported (B= 0.57, t= 11.03, p < .001). The significance tests
(two-tailed and with a normal dispersion) confirmed that the indirect effects were significant
(Sobel, z=3.69, p<0.001). The results of the bootstrap further verified the Sobel test (refer to
Table 6d), with confidence interval (CI) of 95% and the non-zero indirect effects (.04, .12). Thus
Hypothesis 7 received full support for all outcomes. So H1, H3, H5 and H7 were fully
confirmed.
Table 6. Emotional Dissonance as Mediator between Workplace Bullying and Outcomes
a. Job Stress
Variable B SE t p
1 Direct effect of WPB
on Stress
0.21 0.05 4.19 .000
129
2 Direct effects of WPB
on ED
0.27 0.05 5.41 .000
3 Direct effects of ED
on Stress
0.43 0.05 8.20 .000
4 Mediation of ED btw
WPB and Stress
0.33 0.05 6.22 .000
Significance using normal distribution and Indirect effects
Effect SE z p
Sobel 0.12 0.03 4.49 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
ED 0.12 0.03 0.07 0.18
b. Burnout
Variable B SE t p
1 Direct effect of WPB
on Burnout
.39 .06 6.88 .000
2 Direct effects of WPB
on ED
.27 .05 5.41 .000
3 Direct effects of ED
on Burnout
.31 .06 5.22 .000
4 Mediation of ED btw
WPB and Burnout
.47
.06 8.40 .000
130
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .084 .02 3.72 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
ED .08 .03 .04 .14
c. Turnover Intensions
Variable B SE t p
1 Direct effect of WPB
on TOI
.17 .07 2.51 .012
2 Direct effects of WPB
on ED
.27 .05 5.41 .000
3 Direct effects of ED
on TOI
.39 .07 5.41 .000
4 Mediation of ED btw
WPB and TOI
.28 .07 4.06 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .11
.03 3.79 .000
131
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
ED .11 .03 .06 .17
d. Psychological Strain
Variable B SE t p
1 Direct effect of WPB
on Strain
.49 .05 9.51 .000
2 Direct effects of WPB
on ED
.27 .05 5.41 .000
3 Direct effects of ED
on Strain
.28 .05 5.13 .000
4 Mediation of ED btw
WPB and Strain
.57 .05 11.03 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .08 .02 3.69 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
ED .08 .02 .04 .12
132
4.3.2.2 Perceived Organization Support as mediator between Workplace Bullying and Outcomes
(Job Stress, Burnout, Turnover Intensions and Psychological Strain)
Table 7 presents the results of Perceived Organization Support as a mediator between workplace
bullying and outcomes (Job Stress, Burnout, Turnover Intensions and Psychological Strain) and
consequently address H1, H2a, H4a and H6a. The first three hypotheses are related to the direct
effects of bullying on outcomes, then Perceived Organization Support on outcomes, bullying on
Perceived organization support and the last one state the mediation of Perceived Organization
support between bullying and outcomes respectively.
According to Table 7 (a, b, c and d) and in consonance with Hypothesis1 workplace bullying is
associated positively to job stress (B= 0.30, t = 5.59, p < .001), job burnout (B = 0.39, t = 7.05, p
< .001), Turnover Intensions (B = 0.22, t = 3.13, p < .001) and psychological strain (B= 0.52, t =
9.90, p<0.001). Therefore H1 was fully supported. In accordance with Table 7 (a, b, c and d) and
in line with hypothesis 2a Perceived Organization Support was found to be significantly and
negatively associated with job burnout (B= -0.31, t = -5.61, p <0.001), Turnover Intensions (B= -
0.23, t = -3.30, p< 0.001) and psychological strain (B= -0.20, t = -3.83, p < 0.001) and job stress
(B= -0.09, t = -1.73, p< 0.1). Thus, Hypothesis 2a was fully supported for all outcomes except
job stress. Referring to Table 7 (a, b, c and d) and supporting Hypothesis 4a Workplace bullying
was negatively and significantly associated with perceptions of organization support (B= -0.26, t
= -4.83, p < .001). Thus, Hypothesis 4a was fully confirmed.
133
Lastly we tested the mediation hypothesis (Hypothesis 6a) for each outcome, that asserts that
Perceived Organization Support mediates the relationship between workplace bullying and job
stress, which was not supported for job stress (B = 0.33, t = 6.22, p<0.1). The significance tests
(two-tailed and with a normal dispersion) showed that the indirect effects was insignificant
(Sobel, z=1.60, p=n.s). Bootstrap results did not confirm the Sobel test (see Table 7a), with a
bootstrapped 95% CI around the indirect effect which contained zero (-.00, .06).
Next, according to Table 7b, Perceived Organization Support fully acts as a mediator in the
relationship between bullying at the workplace and job burnout which is confirmed (B= 0.47, t =
8.40, p< 0.001). The significance tests (two-tailed and with a normal dispersion) confirmed that
the indirect effects were significant (Sobel, z=3.63, p<0.001). The results of the bootstrap further
verified the Sobel test (refer to Table 7b), with a bootstrap confidence interval (CI) of 95% and
the non-zero indirect effects (.04, .13).
Perceived Organization Support fully mediates the relationship between workplace bullying and
Turnover Intensions (see Table 7c) which is confirmed (B= 0.28, t = 4.06, p < 0.001). The
significance tests (two-tailed and with a normal dispersion) confirmed that the indirect effects
were significant (Sobel, z=2.69, p<0.01). The results of the bootstrap further verified the Sobel
test (refer to Table 7c), with confidence interval (CI) of 95% and the non-zero indirect effects
(.02, .12).
Perceived Organization Support fully mediates the relationship between workplace bullying and
Psychological Strain (see Table 8d) which is supported (B= 0.57, t= 11.03, p< 0.001). The
134
significance tests (two-tailed and with a normal dispersion) confirmed that the indirect effects
were significant (Sobel, z=2.96, p<0.01). The results of the bootstrap further verified the Sobel
test (refer to Table 7d), with confidence interval (CI) of 95% and the non-zero indirect effects
(.02, .09). Thus Hypothesis 6a received full support for all outcomes except Job Stress. So H1
and H4a were fully confirmed whereas H2a and H6a were supported for all outcomes except Job
Stress.
Table 7. Perceived Organizational Support as Mediator between Workplace Bullying and
Outcomes
a. Job Stress
Variable B SE T p
1 Direct effect of WPB
on Stress
.30 .05 5.59 .000
2 Direct effects of WPB
on POS
-.26 .05 -4.83 .000
3 Direct effects of POS
on Stress
-.09 .05 -1.73 .085
4 Mediation of POS btw
WPB and Stress
.33 .05 6.22 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .0239 .0150 1.5957 .1106
135
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
POS .02
.02 -.00 .06
b. Burnout
Variable B SE T p
1 Direct effect of WPB
on Burnout
.39 .06 7.05 .000
2 Direct effects of WPB
on POS
-.26 .06 -4.83 .000
3 Direct effects of POS
on Burnout
-.31 .06 -5.61 .000
4 Mediation of POS btw
WPB and Burnout
.47 .06 8.40 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .08 .02 3.63 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
POS .08 .02 .04 .13
136
c. Turnover Intensions
Variable B SE T p
1 Direct effect of WPB
on TOI
.22 .07 3.13 .001
2 Direct effects of WPB
on POS
-.26 .05 -4.83 .000
3 Direct effects of POS
on TOI
-.23 .07 -3.30 .001
4 Mediation of POS btw
WPB and TOI
.28 .07 4.06 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .06
.02 2.69 .007
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
POS .06 .03 .02 .12
d. Psychological Strain
Variable B SE T p
1 Direct effect of WPB
on Strain
.52 .05 9.90 .000
137
2 Direct effects of WPB
on POS
-.26 .05 -4.83 .000
3 Direct effects of POS
on Strain
-.20 .05 -3.83 .000
4 Mediation of POS btw
WPB and Strain
.57 .05 11.03 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .05
.01 2.96 .003
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
POS .05 .02 .02 .09
4.3.2.3 Perceived Co-worker Support as mediator between Workplace Bullying and Outcomes
(Job Stress, Burnout, Turnover Intensions and Psychological Strain)
Table 8 presents the results of Perceived Co-worker Support as a mediator between workplace
bullying and outcomes (Job Stress, Burnout, Turnover Intensions and Psychological Strain) and
consequently address H1, H2b, H4b and H6b. The first hypothesis states the direct effects of
bullying on outcomes, then the direct effects of perceptions of coworker support on outcomes,
138
bullying on Perceived Co-worker Support and the last one states the mediation of Perceived Co-
worker support between bullying and outcomes respectively.
According to Table 8 (a, b, c and d) and in consonance with Hypothesis1 workplace bullying is
related positively to job stress (B= 0.32, t = 5.85, p < 0.001), job burnout (B= 0.42, t = 7.24, p <
.001), Turnover Intensions (B = 0.31, t = 4.34, p < .001) and psychological strain (B = 0.52, t =
9.65, p < .001). Therefore H1 was fully supported. In accordance with Table 8 (a, b, c and d) and
in line with hypothesis 2b Perceived Co-worker Support was found to be significantly and
negatively associated with job burnout (B = -0.14, t = -2.72, p < .001) and psychological strain
(B = -0.15, t = -3.22, p < .001). However, Perceived Co-worker support was insignificant for job
stress (B = -0.01, t = -.20, p=n.s) and Turnover Intensions (B = 0.10, t = 1.53, p=n.s). Thus,
Hypothesis 2b was supported for all outcomes except job stress and turnover intensions.
Referring to Table 8 (a, b, c and d) and supporting Hypothesis 4b Workplace bullying was
significantly negatively related to Perceived Co-worker Support (B = -0.35, t = -5.87, p < .001).
Thus, Hypothesis 4b was fully confirmed.
Lastly we tested the mediation hypothesis (Hypothesis 6b) for each outcome, which asserts that
perceptions of Co-worker Support acts as a mediator in the relationship between workplace
bullying and job stress, which was not supported for job stress (B = 0.33, t = 6.22, p<0.1). The
significance tests (two-tailed and with a normal dispersion) depicted that the indirect effects were
insignificant (Sobel, z=0.19, p=n.s.). The results of the bootstrap further verified the Sobel test
(refer to Table 8a), with confidence interval (CI) of 95% and the indirect effects containing zero
(-.03, .04).
139
Next, according to Table 8b, Perceived Co-worker Support fully mediates the relationship
between workplace bullying and job burnout which is supported (B = 0.47, t = 8.40, p < .001).
The significance tests (two-tailed and with a normal dispersion) confirmed that the indirect
effects were significant (Sobel, z=2.44, p<0.05). The results of the bootstrap further verified the
Sobel test (refer to Table 8b), with confidence interval (CI) of 95% and the non-zero indirect
effects (.02, .09).
Perceived Co-worker Support did not mediate the relationship between workplace bullying and
Turnover Intensions (see Table 8c) (B = 0.28, t = 4.06, p=n.s). The significance tests (two-tailed
and with a normal dispersion) showed that the indirect effects was insignificant (Sobel, z=-1.46,
p=n.s.). Bootstrap results did not confirm the Sobel test (see Table 8c), with a bootstrapped 95%
CI around the indirect effect which contained zero (-.08, .01).
Perceived Co-worker Support fully mediates the relationship between workplace bullying and
Psychological Strain (see Table 8d) which is supported (B= 0.57, t= 11.03, p < 0.001). The
significance tests (two-tailed and with a normal dispersion) confirmed that the indirect effects
were significant (Sobel, z=2.80, p<0.01). The results of the bootstrap further verified the Sobel
test (refer to Table 8d), with confidence interval (CI) of 95% and the non-zero indirect effects
(.02, .10). Thus Hypothesis 6b received full support for all outcomes except Job Stress and
Turnover Intensions. So H1 and H4b were fully confirmed whereas H2b and H6b were
supported for all outcomes except Job Stress and turnover Intensions.
140
Table 8. Perceived Co-worker Support as Mediator between Workplace Bullying and
Outcomes
a. Job Stress
Variable B SE t p
1 Direct effect of WPB
on Stress
.32 .06 5.85 .000
2 Direct effects of WPB
on PCS
-.35 .06 -5.87 .000
3 Direct effects of PCS
on Stress
-.01 .05 -.20 .844
4 Mediation of PCS btw
WPB and Stress
.33 .05 6.22 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .00 .02 .19 .846
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PCS .00 .02 -.03 .04
b. Burnout
Variable B SE t p
1 Direct effect of WPB
on Burnout
.42 .06 7.24 .000
141
2 Direct effects of WPB
on PCS
-.35 .06 -5.87 .000
3 Direct effects of PCS
on Burnout
-.14 .05 -2.72 .006
4 Mediation of PCS btw
WPB and Burnout
.47 .06 8.40 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .05 .02 2.44 .014
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PCS .05 .02 .02 .09
c. Turnover Intensions
Variable B SE t p
1 Direct effect of WPB
on TOI
.31 .07 4.34 .000
2 Direct effects of WPB
on PCS
-.35 .06 -5.87 .000
3 Direct effects of PCS
on TOI
.10 .06 1.53 .127
4 Mediation of PCS btw
WPB and TOI
.28 .07 4.06 .000
142
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel -.03 .02 -1.46 .14
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PCS -.03 .02 -.08 .01
d. Psychological Strain
Variable B SE t p
1 Direct effect of WPB
on Strain
.52 .05 9.65 .000
2 Direct effects of WPB
on PCS
-.35 .06 -5.87 .000
3 Direct effects of PCS
on Strain
-.15 .05 -3.22 .001
4 Mediation of PCS btw
WPB and Strain
.57 .05 11.03 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .05 .02 2.80 .005
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
143
Effect
PCS .05 .02 .02 .10
4.3.2.4 Perceived Supervisor Support as mediator between Workplace Bullying and Outcomes
(Job Stress, Burnout, Turnover Intensions and Psychological Strain)
Table 9 presents the results of Perceived Supervisor Support as a mediator between workplace
bullying and outcomes (Job Stress, Burnout, Turnover Intensions and Psychological Strain) and
consequently address H1, H2c, H4c and H6c. The first hypothesis states the direct effects of
bullying on outcomes, then the direct effects of Perceived Supervisor Support on outcomes,
bullying on Perceived Supervisor Support and the last one states the mediation of Perceived
Supervisor support between bullying and outcomes respectively.
According to Table 9 (a, b, c and d) and in consonance with Hypothesis1 workplace bullying is
associated positively with job stress (B= 0.33, t = 5.83, p < .001), job burnout (B= 0.38, t = 6.43,
p<0.001), Turnover Intensions (B= 0.25, t = 3.37, p < 0.001) and psychological strain (B= 0.50,
t = 9.12, p < 0.001). Therefore H1 was fully supported. In accordance with Table 9 (a, b, c and d)
and in line with hypothesis 2c Perceived Supervisor Support was established to be significantly
and negatively related to job burnout (B= -0.27, t= -4.63, p < 0.001) and psychological strain
(B= -0.21, t = -3.65, p < .001). However, Perceived Supervisor support was insignificant for job
stress (B =0.01, t =.15, p=n.s) and Turnover Intensions (B =-0.09, t =-1.12, p=n.s). Thus,
Hypothesis 2c was supported for all outcomes except job stress and turnover intensions.
Referring to Table 9 (a, b, c and d) and supporting Hypothesis 4c Workplace bullying was
144
negatively and significantly associated to perceived supervisor support (B = -0.35, t = -7.16, p <
.001). Thus, Hypothesis 4c was fully confirmed.
Lastly we tested the mediation hypothesis (Hypothesis 6c) for each outcome, that contends that
perceptions of supervisor support mediates the relationship between workplace bullying and job
stress, which was not supported for job stress (B= 0.33, t = 6.22, p<0.1). The significance tests
(two-tailed and with a normal dispersion) depicted that the indirect effects was insignificant
(Sobel, z=-0.15, p=n.s.). Bootstrap results did not confirm the Sobel test (see Table 9a), with a
bootstrapped 95% CI around the indirect effect which contained zero (-.05, .04).
Next, according to Table 9b, perceptions of supervisor support fully mediates the relationship
between workplace bullying and job burnout which is confirmed (B= 0.47, t = 8.40, p < .001).
The significance tests (two-tailed and with a normal dispersion) confirmed that the indirect
effects were significant (Sobel, z=3.69, p<0.001). The results of the bootstrap further verified the
Sobel test (refer to Table 9b), with bootstrapped confidence interval (CI) of 95% and the non-
zero indirect effects (.05, .15).
Perceived Supervisor Support did not mediate the relationship between workplace bullying and
Turnover Intensions (see Table 9c) (B = 0.28, t = 4.06, p=n.s). The significance tests (two-tailed
and with a normal dispersion) highlighted that the indirect effects were insignificant (Sobel,
z=1.09, p=n.s.). The results of the did not confirm the Sobel test (see Table 9c), with a
bootstrapped 95% CI around the indirect effect which contained zero (-.02, .09).
145
Perceptions of supervisory support fully mediates the relationship between workplace bullying
and Psychological Strain (see Table 9d) which is supported (B= 0.57, t = 11.03, p < .001). The
significance tests (two-tailed and with a normal dispersion) confirmed that the indirect effects
were significant (Sobel, z=3.22, p<0.001). The results of the bootstrap further verified the Sobel
test (refer to Table 9d), with confidence interval (CI) of 95% and the non-zero indirect effects
(.04, .12). Thus Hypothesis 6c received full support for all outcomes except Job Stress and
Turnover Intensions. So H1 and H4c were fully confirmed whereas H2c and H6c were supported
for all outcomes except Job Stress and turnover Intensions.
Table 9. Perceived Supervisor Support as Mediator between Workplace Bullying and
Outcomes
a. Job Stress
Variable B SE t p
1 Direct effect of WPB
on Stress
.33 .06 5.83 .000
2 Direct effects of WPB
on PSS
-.35 .05 -7.16 .000
3 Direct effects of PSS
on Stress
.01 .06 .15 .879
4 Mediation of PSS btw
WPB and Stress
.33 .05 6.22 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel -.00 .02 -.15 .88
146
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PSS -.00 .02 -.05 .04
b. Burnout
Variable B SE t p
1 Direct effect of WPB
on Burnout
.38 .06 6.43 .000
2 Direct effects of WPB
on PSS
-.35 .05 -7.16 .000
3 Direct effects of PSS
on Burnout
-.27 .06 -4.36 .000
4 Mediation of PSS btw
WPB and Burnout
.47 .06 8.40 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .09 .03 3.69 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PSS .09
.02 .05 .15
147
c. Turnover Intensions
Variable B SE t p
1 Direct effect of WPB
on TOI
.25 .07 3.37 .000
2 Direct effects of WPB
on PSS
-.35 .05 -7.16 .000
3 Direct effects of PSS
on TOI
-.09 .08 -1.12 .264
4 Mediation of PSS btw
WPB and TOI
.28 .07 4.06 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .03 .03 1.09 .274
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PSS .03 .03 -.02 .09
d. Psychological Strain
Variable B SE t p
1 Direct effect of WPB
on Strain
.50 .05 9.12 .000
2 Direct effects of WPB
on PSS
-.35 .05 -7.16 .000
148
3 Direct effects of PSS
on Strain
-.21 .06 -3.65 .000
4 Mediation of PSS btw
WPB and Strain
.57 .05 11.03 .000
Indirect Effect and Significance using normal distribution
Effect SE Z p
Sobel .07 .02 3.22 .001
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PSS .07 .02 .04 .12
4.3.3 Regression Results for Parallel and Sequential Mediation using Bootstrapping
4.3.3.1 Emotional Dissonance and Perceived Supervisor Support as Parallel and Sequential
mediators between Workplace Bullying and Outcomes
Table 10 presents the results of Perceived Supervisor Support and Emotional Dissonance as
parallel and sequential mediators between workplace bullying and outcomes (Job Stress,
Burnout, Turnover Intensions and Psychological Strain) and consequently address H6c, H7, H8.
The first hypothesis (i.e. H6c) relates to the parallel path or in other words simple mediating
effect of Perceived Supervisor Support as mediator between bullying and outcomes, the second
hypothesis (i.e. H7) relates to the other parallel path or in other words simple mediating effect of
Emotional Dissonance as mediator between bullying and outcomes and the last one deals with
149
the sequential mediation of Perceived Supervisor support and Emotional Dissonance between
bullying and outcomes respectively.
According to Table 10 (a, b, c and d) (Hypothesis 6c) revealed for each outcome that Perceptions
of supervisor support mediates the relationship between workplace bullying and job burnout with
a confidence interval of 95% bootstrapped and which contains non-zero around the indirect
effects (.06, .16)(see Table 10b). Moreover, Perceived Supervisor Support mediates the
relationship between workplace bullying and Psychological Strain with a confidence interval of
95% bootstrapped and which contains non-zero around the indirect effects (.05, .14)(refer to
Table 10c). However, Perceived Supervisor Support did not mediate the relationship between
workplace bullying and job stress as bootstrap results for indirect effects were insignificant (-.02,
.06)(see Table 10a) and for turnover intensions as well (-.01, .11) )(see Table 10d). Thus, the
parallel or simple mediating effects of perceived supervisor support as a mediator between
workplace bullying was supported for all outcomes except job stress and turnover intensions.
Thus H6c was confirmed for all outcomes except Job Stress and Turnover Intensions.
According to Table 10(a, b, c and d) (Hypothesis 7) revealed for each outcome that Emotional
Dissonance fully mediated the relationship between workplace bullying and job stress with a
confidence interval of 95% bootstrapped and which contains non-zero around the indirect effects
(.08, .20)(refer to Table 10a), Job burnout (.06, .17) )(see Table 10b), Psychological Strain (.05,
.15)(see Table 10c) and Turnover Intensions (.07, .20) )(see Table 10d). Thus, the parallel or
simple mediating effects of Emotional Dissonance as a mediator between workplace bullying
was fully supported. Thus H7 was fully confirmed.
150
According to Table 10(a, b, c and d) (Hypothesis 8) revealed for each outcome that Perceived
Supervisor Support and Emotional Dissonance sequentially mediated the relationship between
workplace bullying and job stress with a confidence interval of 95% bootstrapped and which
contains non-zero around the indirect effects (-.04, -.00)(see Table 10a), Job burnout (-.03,-.00)
(see Table 10b), Psychological Strain (-.03, -.00)(see Table 10c) and Turnover Intensions (-.04, -
.00) )(see Table 10d). Thus, the sequential mediation effects of Perceived Supervisor Support
and Emotional Dissonance as a mediator between workplace bullying was fully supported. Thus
H8 was fully confirmed. Thus, H7 was fully supported, H6c was supported for all outcomes
except job stress and turnover intensions and H8 was fully confirmed for Perceived Supervisor
support.
Table 10. Perceived Supervisor Support and Emotional Dissonance as parallel and
sequential mediators between Workplace Bullying and Outcomes
a. Job Stress
Variable B SE t p
1 Direct effect of WPB
on PSS
-.35 .05 -7.16 .000
2 Direct effects of WPB
on ED
.31 .05 5.82 .000
3 Direct effects of PSS
on ED
.12 .06 2.08 .038
4 Direct effects of WPB .19 .05 3.58 .000
151
on Stress
5 Direct effects of PSS
on Stress
-.04 .05 -.77 .441
6 Direct effects of ED
on Stress
.43 .05 8.23 .000
7 Mediation of PSS and
ED btw WPB and
Stress
.33 .05 6.22 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PSS .01 .02 -.02 .06
PSS and ED -.02 .01 -.04 -.00
ED .14
.03 .08 .20
b. Burnout
Variable B SE t p
1 Direct effect of WPB
on PSS
-.35 .05 -7.16 .000
2 Direct effects of WPB
on ED
.31 .05 5.82 .000
3 Direct effects of PSS
on ED
.12 .06 2.08 .038
152
4 Direct effects of WPB
on Burnout
.27 .06 4.61 .000
5 Direct effects of PSS
on Burnout
-.31 .06 -5.23 .000
6 Direct effects of ED
on Burnout
.34 .06 5.99 .000
7 Mediation of PSS and
ED btw WPB and
Burnout
.47 .06 8.40 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PSS .11 .02 .06 .16
PSS and ED -.01 .01 -.03 -.00
ED .11 .03 .06 .17
c. Psychological Strain
Variable B SE t p
1 Direct effect of WPB
on PSS
-.35 .05 -7.16 .000
2 Direct effects of WPB
on ED
.31 .05 5.82 .000
3 Direct effects of PSS
on ED
.12 .06 2.08 .038
4 Direct effects of WPB .40 .05 7.34 .000
153
on Strain
5 Direct effects of PSS
on Strain
-.24 .05 -4.45 .000
6 Direct effects of ED
on Strain
.31 .05 5.75 .000
7 Mediation of PSS and
ED btw WPB and
Strain
.57 .05 11.03 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PSS .09 .02 .05 .14
PSS and ED -.01 .01 -.03 -.00
ED .10 .02 .05 .15
d. Turnover Intensions
Variable B SE t p
1 Direct effect of WPB
on PSS
-.35 .05 -7.16 .000
2 Direct effects of WPB
on ED
.31 .05 5.82 .000
3 Direct effects of PSS
on ED
.12 .06 2.08 .038
4 Direct effects of WPB
on TOI
.12 .07 1.64 .101
154
5 Direct effects of PSS
on TOI
-.13 .07 -1.80 .073
6 Direct effects of ED
on TOI
.41 .07 5.59 .000
7 Mediation of PSS and
ED btw WPB and TOI
.28 .07 4.06 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PSS .05 .03 -.00 .11
PSS and ED -.02 .01 -.04 -.00
ED .13 .03 .07 .20
4.3.3.2 Emotional Dissonance and Perceived Co-worker Support as Parallel and Sequential
mediators between Workplace Bullying and Outcomes
Table 11 presents the results of Perceived Co-worker Support and Emotional Dissonance as
parallel and sequential mediators between workplace bullying and outcomes (Job Stress,
Burnout, Turnover Intensions and Psychological Strain) and consequently address H6b, H7, H8.
The first hypothesis (i.e. H6b) relates to the parallel path or in other words simple mediating
effect of Perceived Co-worker Support as mediator between bullying and outcomes, the second
hypothesis (i.e. H7) relates to the other parallel path or in other words simple mediating effect of
Emotional Dissonance as mediator between bullying and outcomes and the last one deals with
155
the sequential mediation of Perceived Co-worker support and Emotional Dissonance between
bullying and outcomes respectively.
According to Table 11(a, b, c and d) (Hypothesis 6b) revealed for each outcome that Perceived
Co-worker Support mediates the association between bullying at the workplace and job burnout
with a confidence interval of 95% bootstrapped and which contains non-zero around the indirect
effects (.03, .11)(refer to Table 11b). Moreover, Perceived Co-worker Support mediates the
relationship between workplace bullying and Psychological Strain with a confidence interval of
95% bootstrapped and which contains non-zero around the indirect effects (.03, .12)(refer to
Table 11c). However, Perceived Co-worker Support did not mediate the relationship between
workplace bullying and job stress as bootstrap results for indirect effects were insignificant (-.00,
.06) (see Table 11a) and for turnover intensions as well (-.06, .02)(see Table 11d). Thus, the
parallel or simple mediating effects of perceived co-worker support as a mediator between
workplace bullying was supported for all outcomes except job stress and turnover intensions.
Thus H6b was confirmed for all outcomes except Job Stress and Turnover Intensions.
According to Table 11(a, b, c and d) (Hypothesis 7) revealed for each outcome that Emotional
Dissonance fully mediated the relationship between workplace bullying and job stress with a
confidence interval of 95% bootstrapped and which contains non-zero around the indirect effects
(.08, .21)(refer to Table 11a), Job burnout (.06, .17) )(see Table 11b), Psychological Strain (.06,
.16)(see Table 11c) and Turnover Intensions (.07, .19) )(see Table 11d). Thus, the parallel or
simple mediation effects of Emotional Dissonance as a mediator between workplace bullying
and outcomes was fully supported. Thus H7 was fully confirmed.
156
According to Table 11(a, b, c and d) (Hypothesis 8) revealed for each outcome that Perceived
Co-worker Support and Emotional Dissonance sequentially mediated the relationship between
workplace bullying and job stress with a confidence interval of 95% bootstrapped and which
contains non-zero around the indirect effects (-.04, -.01)(refer to Table 11a), Job burnout (-.04,-
.01)(see Table 11b), Psychological Strain (-.04, -.01)(see Table 11c) and Turnover Intensions (-
.04, -.01) )(see Table 11d). Thus, the sequential mediation effects of Perceived Co-worker
Support and Emotional Dissonance as a mediator between workplace bullying and outcomes was
fully supported. Thus H8 was fully confirmed. Thus, H7 was fully confirmed, H6b was
supported for all outcomes except job stress and turnover intensions and H8 was fully supported
for Perceived Co-worker support as sequential mediator between workplace bullying and
outcomes.
Table 11. Perceived Co-worker Support and Emotional Dissonance as parallel and
sequential mediators between Workplace Bullying and Outcomes
a. Job Stress
Variable B SE t p
1 Direct effect of WPB
on PCS
-.35 .06 -5.87 .000
2 Direct effects of WPB
on ED
.32 .05 6.18 .000
3 Direct effects of PCS
on ED
.14 .05 3.14 .001
4 Direct effects of WPB .18 .05 3.39 .000
157
on Stress
5 Direct effects of PCS
on Stress
-.07 .04 -1.64 .101
6 Direct effects of ED
on Stress
.45 .05 8.38 .000
7 Mediation of PCS and
ED btw WPB and
Stress
.33 .05 6.22 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PCS .03 .02 -.00 .06
PCS and ED -.02 .01 -.05 -.01
ED .14 .03 .09 .21
b. Burnout
Variable B SE t p
1 Direct effect of WPB
on PCS
-.35 .06 -5.87 .000
2 Direct effects of WPB
on ED
.32 .05 6.18 .000
3 Direct effects of PCS
on ED
.14 .05 3.14 .001
4 Direct effects of WPB
on Burnout
.31 .06 5.29 .000
158
5 Direct effects of PCS
on Burnout
-.19 .05 -3.82 .000
6 Direct effects of ED
on Burnout
.35 .06 5.90 .000
7 Mediation of PCS and
ED btw WPB and
Burnout
.47 .06 8.40 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PCS .07 .02 .03 .11
PCS and ED -.02
.01 -.04 -.01
ED .11 .03 .06 .17
c. Psychological Strain
Variable B SE t p
1 Direct effect of WPB
on PCS
-.35 .06 -5.87 .000
2 Direct effects of WPB
on ED
.32 .05 6.18 .000
3 Direct effects of PCS
on ED
.14 .05 3.14 .001
4 Direct effects of WPB .41 .05 7.69 .000
159
on Strain
5 Direct effects of PCS
on Strain
-.20 .05 -4.35 .000
6 Direct effects of ED
on Strain
.32 .05 5.93 .000
7 Mediation of PCS and
ED btw WPB and
Strain
.57 .05 11.03 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PCS .07 .02 .03 .12
PCS and ED -.02 .01 -.04 -.01
ED .10 .03 .06 .16
d. Turnover Intensions
Variable B SE t p
1 Direct effect of WPB
on PCS
-.35 .06 -5.87 .000
2 Direct effects of WPB
on ED
.32 .05 6.18 .000
3 Direct effects of PCS
on ED
.14 .05 3.14 .001
160
4 Direct effects of WPB
on TOI
.19 .07 2.59 .010
5 Direct effects of PCS
on TOI
.04 .06 .68 .498
6 Direct effects of ED
on TOI
.39 .07 5.21 .000
7 Mediation of PCS and
ED btw WPB and TOI
.28 .07 4.06 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
PCS -.01 .02 -.06 .03
PCS and ED -.01 .01 -.04 -.01
ED .12 .03 .07 .19
4.3.3.3 Emotional Dissonance and Perceived Organization Support as Parallel and Sequential
mediators between Workplace Bullying and Outcomes
Table 12 presents the results of Perceived Organization Support and Emotional Dissonance as
parallel and sequential mediators between workplace bullying and outcomes (Job Stress,
Burnout, Turnover Intensions and Psychological Strain) and consequently address H6a, H7, H8.
The first hypothesis (i.e. H6a) relates to the parallel path or in other words simple mediating
effect of Perceived Organization Support as mediator between bullying and outcomes, the
161
second hypothesis (i.e. H7) relates to the other parallel path or in other words simple mediating
effect of Emotional Dissonance as mediator between bullying and outcomes and the last one
deals with the sequential mediation of Perceived Organization support and Emotional
Dissonance between bullying and outcomes respectively.
According to Table 12(a, b, c and d) (Hypothesis 6a) revealed for each outcome that Perceived
Organization Support mediates the relationship between workplace bullying and job burnout
with a confidence interval of 95% bootstrapped and which contains non-zero around the indirect
effects (.04, .12)(refer to Table 12b). Moreover, Perceived Organization Support mediates the
relationship between workplace bullying and Psychological Strain with a confidence interval of
95% bootstrapped and which contains non-zero around the indirect effects (.02, .09)(refer to
Table 12c) and turnover intensions (.02, .11) )(see Table 12d). However, Perceived Organization
Support did not mediate the relationship between workplace bullying and job stress as bootstrap
results for indirect effects were insignificant (-.01, .05) (see Table 12a). Thus, the parallel or
simple mediating effects of perceived organization support as a mediator between workplace
bullying was supported for all outcomes except job stress. Thus H6a was verified for all the
outcomes except Job Stress.
According to Table 12(a, b, c and d) (Hypothesis 7) revealed for each outcome that Emotional
Dissonance fully mediated the relationship between workplace bullying and job stress with a
confidence interval of 95% bootstrapped and which contains non-zero around the indirect effects
(.06, .17)(refer to Table 12a), Job burnout (.04, .13)(see Table 12b), Psychological Strain (.04,
.12)(see Table 12c) and Turnover Intensions (.05, .16) )(see Table 12d). Thus, the parallel or
162
simple mediation effects of Emotional Dissonance as a mediator between workplace bullying
and outcomes was fully supported. Thus H7 was fully confirmed.
According to Table 12(a, b, c and d) (Hypothesis 8) revealed for each outcome that Perceived
Organization Support and Emotional Dissonance did not sequentially mediate the relationship
between workplace bullying and job stress with a bootstrapped 95% CI around the indirect effect
not containing zero (-.01,.02)(see Table 12a), Job burnout (-.00, .02)(see Table 12b),
Psychological Strain (-.00, .01)(see Table 12c) and Turnover Intensions (-.00, .02)(see Table
12d). Thus, the sequential mediation effects of Perceived Organization Support and Emotional
Dissonance as a mediator between workplace bullying was rejected. Thus H8 was not supported
for Perceived Organization Support and Emotional Dissonance as Sequential mediator between
workplace bullying and outcomes. Thus, simple mediation or parallel H7 was fully supported
H6a was supported for all outcomes except job stress whereas H8 was not confirmed for POS
only.
Table 12. Perceived Organizational Support and Emotional Dissonance as parallel and
sequential mediators between Workplace Bullying and Outcomes
a. Job Stress
Variable B SE t p
1 Direct effect of WPB
on POS
-.26 .05 -4.83 .000
2 Direct effects of WPB .26 .05 4.95 .000
163
on ED
3 Direct effects of POS
on ED
-.06 .05 -1.07 .285
4 Direct effects of WPB
on Stress
.19 .05 3.76 .000
5 Direct effects of POS
on Stress
-.07 .05 -1.41 .160
6 Direct effects of ED
on Stress
.43 .05 8.11 .000
7 Mediation of POS and
ED btw WPB and
Stress
.33 .05 6.22 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
POS .02 .01 -.01 .05
POS and ED .01 .01 -.01 .02
ED .11 .03 .06 .17
b. Burnout
Variable B SE t p
1 Direct effect of WPB
on POS
-.26 .05 -4.83 .000
2 Direct effects of WPB .26 .05 4.95 .000
164
on ED
3 Direct effects of POS
on ED
-.06 .05 -1.07 .285
4 Direct effects of WPB
on Burnout
.32 .06 5.71 .000
5 Direct effects of POS
on Burnout
-.30 .05 -5.51 .000
6 Direct effects of ED
on Burnout
.29 .06 5.11 .000
7 Mediation of POS and
ED btw WPB and
Burnout
.47 .06 8.40 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
POS .08 .02 .04 .12
POS and ED .00
.00 -.00 .02
ED .07 .02 .04 .13
c. Psychological Strain
Variable B SE t p
1 Direct effect of WPB
on POS
-.26 .05 -4.83 .000
2 Direct effects of WPB .26 .05 4.95 .000
165
on ED
3 Direct effects of POS
on ED
-.06 .05 -1.07 .285
4 Direct effects of WPB
on Strain
.45 .05 8.57 .000
5 Direct effects of POS
on Strain
-.19 .05 -3.67 .000
6 Direct effects of ED
on Strain
.27 .05 5.00 .000
7 Mediation of POS and
ED btw WPB and
Strain
.57 .05 11.03 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
POS .05 .02 .02 .09
POS and ED .00 .00 -.00 .01
ED .07 .02 .04 .12
d. Turnover Intensions
Variable B SE t p
1 Direct effect of WPB
on POS
-.26 .05 -4.83 .000
2 Direct effects of WPB
on ED
.26 .05 4.95 .000
166
3 Direct effects of POS
on ED
-.06 .05 -1.07 .286
4 Direct effects of WPB
on TOI
.12 .07 1.75 .081
5 Direct effects of POS
on TOI
-.21 .07 -3.12 .002
6 Direct effects of ED
on TOI
.38 .07 5.28 .000
7 Mediation of POS and
ED btw WPB and TOI
.28 .07 4.06 .000
Bootstrap results for indirect effects
M SE LL95% CI UL 95% CI
Effect
POS .05 .02 .02 .11
POS and ED .01 .01 -.00 .02
ED .10 .03 .05 .16
4.3.4 Regression Results for Moderation using Bootstrapping
4.3.4.1 Tests of Moderation
To test moderation hypothesis, we employed the macros devised by Preacher and Hayes (2004).
The process technique offered by these authors deviates from the Baron and Kenny's traditional
approach of conducting moderation analysis i-e to test moderation using normal regression the
interaction term is created comprising the IV and the moderator variable. The macro developed
167
by these authors allows the user to directly enter the moderator variable without the need of
creating the cross product term of the IV and moderator variable. In addition, in the options
section, variables are automatically centered by checking the mean center for products box.
Moderation takes place when the relationship between the predictor and the outcome variable is
contingent on a third factor which enhances the strength of the relationship between the IV and
the DV. This third factor or moderator variable in combination with the independent variable
influences the outcome variable (Preacher & Hayes, 2004; Preacher, Rucker & Hayes, 2007).
If the cross product term of IV and moderator is significant, the interactive effects may be
checked to establish if the slope of X on Y is significant for given values of the moderator (W)
i.e. normally it is plotted one standard deviation above and below the mean point. This technique
is explained at greater length by Aiken and West (1991). The simple regression of dependant
variable (Y) on independent variable (X) at different value points of moderator (W) are also
drawn to help explanation. Instead of deciding on a fewer amount of random values of moderator
(W) and examining the p-values of the simple slopes at those points, we can look for the points
of moderator (W) for which the slope of Y is significant on X.
4.3.4.2 Core Self Evaluation as a Moderator between Workplace Bullying and Emotional
Dissonance
Hypothesis 10 states that core self evaluation will moderate the relationship between Workplace
Bullying and Emotional Dissonance which was supported (B= 0.13, t = 2.26, p < .05). The
significance tests (two-tailed and with a normal dispersion) confirmed that the indirect effects
168
were significant (Effect = .42, Boot SE= .08, t = 5.00, p < .001). The results of the bootstrap
further verified these effects (refer to Table 13a), with confidence interval (CI) of 95% and the
non-zero indirect effects (.25, .58).
Figure 16. Interaction effects of Workplace Bullying and Core Self Evaluations on
Emotional Dissonance
Figure 16 depicts the interaction plot of workplace bullying x Core self evaluation for emotional
dissonance. Contrary to expectations, the slope of the relationship between workplace bullying
and emotional dissonance was relatively negative for negative core self evaluations (simple slope
= -.44, p < .05), whereas the slope was weak for positive core self evaluations (simple slope =-
.20, p=n.s). Thus, hypothesis 10 was partially supported.
Low High
Workplace Bullying
Em
oti
onal
Dis
sonan
ce
169
4.3.4.3 Core Self Evaluation as a Moderator between Workplace Bullying and Perceived
Supervisor Support
Hypothesis 9c states that core self evaluation will moderate the relationship between Workplace
Bullying and perceptions of supervisory support which was significant (B=-0.15, t= -2.60, p <
.01). The significance tests (two-tailed and with a normal dispersion) confirmed that the indirect
effects were significant (Effect = -.44, Boot SE= .08, t = -5.27, p < .001). The results of the
bootstrap further verified these effects (refer to Table 13b), with confidence interval (CI) of 95%
and the non-zero indirect effects (-.60, -.28).
Figure 17 shows the interaction plot of workplace bullying x Core self evaluation for perceived
supervisor support. Contrary to expectations, the slope of the relationship between workplace
bullying and emotional exhaustion was positive and strong for negative core self evaluations
(simple slope = .55, p < .01), whereas the slope was weak for positive core self evaluations
(simple slope =.43, t = 6.84, p<.01). Thus, hypothesis 9c was partially supported.
Low High
Workplace Bullying
Per
ceiv
ed S
uper
vis
or
Support
170
Figure 17. Interaction effects of Workplace Bullying and Core Self Evaluations on
Perceived Supervisor Support
4.3.4.4 Core Self Evaluation as a Moderator between Workplace Bullying and Perceived Co-
worker Support
Hypothesis 9b states that core self evaluation moderates the relationship between Workplace
Bullying and Perceived Co-worker Support which was not supported (B =-0.09, t = -1.50,
p=n.s). The significance tests (two-tailed and with a normal dispersion) depicted that the indirect
effects were insignificant (Effect = -.09, Boot SE= .07, t = -1.67, p=n.s.). The results of the
bootstrap highlighted no effects (refer to Table 13c), with confidence interval (CI) of 95% and
indirect effect containing zero (-.31, .02). Thus, Hypothesis 9b was rejected.
4.3.4.5 Core Self Evaluation as a Moderator between Workplace Bullying and Perceived
Organization Support
Hypothesis 9a states that core self evaluations moderates the relationship between Workplace
Bullying and perceptions of organization support which was rejected (B =-0.08, t = -1.07,
p=n.s). The significance tests (two-tailed and with a normal dispersion) showed that the indirect
effects were insignificant (Effect = -.09, Boot SE= .07, t = -1.32, p=n.s.). The results of the
bootstrap highlighted these insignificant effects (refer to Table 13d), with confidence interval
(CI) of 95% and the non-zero indirect effects (-.22, .04).
171
Thus, Hypothesis 9a was rejected. Table 14 also provides details of the moderation analysis
through the regression method in which step-1 shows the controls, step-2 consists of independent
variable of workplace bullying and moderator core self evaluations whereas step-3 depicts the
interaction of workplace bullying x core self evaluations for all the four outcomes i.e. Emotional
Dissonance, perceptions of coworker support, perceptions of supervisory support and Perceived
Organization support.
Table 13. Results of Moderated Regression Analyses for Core Self Evaluations as
Moderator in the relationship between Workplace Bullying and Outcomes
a. Emotional Dissonance (Time-3)
Predictor B SE t p
Constant 3.43 .26 13.36 .000
CSE (Time-2) -.17 .07 -2.32 .020
WB (Time-1) .30 .05 5.44 .000
WB*CSE .13 .06 2.26 .024
Conditional direct effects of X on Y
CSE (Time-2) Effect Boot SE T p LLCI ULCI
CSE -1 SD(-.93) .17 .07 2.46 .01 .03 .31
CSE M (.00) .30
.05 5.44 .00 .19 .40
CSE +1 SD (.93) .42 .08 5.00 .00 .25 .58
b. Perceived Supervisor Support (Time-3)
Predictor B SE t p
Constant 4.99 .26 19.44 .000
CSE (Time-2) .30 .07 4.15 .000
172
WB (Time-1) -.30 .05 -5.52 .000
WB*CSE -.15 .06 -2.60 .009
Conditional direct effects of X on Y
CSE (Time-2) Effect Boot SE T p LLCI ULCI
CSE -1 SD(-.93) -.16 .07 -2.26 .02 -.29 -.02
CSE M (.00) -.30 .05 -5.52 .00 -.40 -.19
CSE +1 SD (.93) -.44 .08 -5.27 .00 -.60 -.28
c. Perceived Co-worker Support (Time-3)
Predictor B SE t p
Constant 4.82 .31 15.40 .000
CSE (Time-2) .34 .09 3.93 .000
WB (Time-1) -.21 .07 -3.22 .001
WB*CSE -.08 .07 -1.07 .287
Conditional direct effects of X on Y
CSE (Time-2) Effect Boot SE T p LLCI ULCI
CSE -1 SD(-.93) -.14
.08 -1.67 .10 -.31 .025
CSE M (.00) . 21 .07 -3.22 .00
-.34 -.082
CSE +1 SD (.93) -.28 .10 -2.78 .01 -.48 -.083
d. Perceived Organization Support (Time-3)
Predictor B SE t p
Constant 4.63 .25 18.34 .000
CSE (Time-2) .48 .07 6.83 .000
WB (Time-1) -.17 .05 -3.21 .002
WB*CSE -.09 .06 -1.50 .134
Conditional direct effects of X on Y
CSE (Time-2) Effect Boot SE T p LLCI ULCI
CSE -1 SD(-.93) -.09
.07 -1.32 .19 -.22 .04
173
CSE M () -.17 .05 -3.21 .00 -.28 -.07
CSE +1 SD (.93) -.25
.08 -3.06 .00 -.41 -.09
4.3.5 Moderated Regression Analysis
Table 14. Results of Moderated Regression Analyses
Dependant Variables at Time-3
Emotional
Dissonance
β
POS
β
PCS
β
PSS
β
Step 1
Age .14* -.01 -.02 -.05
Gender .12* .02 -.01 .01
Step ∆R2 .03* .00 .00 .00
Step 2
Workplace Bullying (WPB-T1) .28*** -.16** -.17** -.27***
Core Self Evaluations (CSE-T2) -.16** .38*** .24*** .25***
Step ∆R2 .13*** .20*** .11*** .18***
Step 3
WPBxCSE .58* -.37 -.28 -.66*
Step ∆R2 .01* .01 .00 .02*
Note. N=334; for Gender 0= ‘Male’ and 1= ‘Female’; all predictors and interactions were mean-
centered.
†p< .10, *p< .05, **p< .01, ***p< .001
174
Table 15. Summary of Findings of proposed Hypothesis
a. Direct Effect Results
Sr
No
Hyp
No
IV DV Result for each
outcome
Overall Result
1 H1 Workplace
Bullying
Job Stress Confirmed
Fully
Confirmed
Job Burnout Confirmed
Turnover Intensions Confirmed
Psychological Strain Confirmed
2 H2a
Perceived
Organization
Support
Job Stress Not Confirmed (ns)
Confirmed for
all outcomes
except Job
Stress
Job Burnout Confirmed
Turnover Intensions Confirmed
Psychological Strain Confirmed
3 H2b Perceived Co-
worker Support
Job Stress Not Confirmed (ns)
Fully
Confirmed for
Burnout and
Strain only
Job Burnout Confirmed
Psychological Strain Confirmed
Turnover Intensions Not Confirmed (ns)
4 H2c
Perceived
Supervisor
Support
Job Stress Not Confirmed (ns)
Fully
Confirmed for
Burnout and
Strain only
Job Burnout Confirmed
Turnover Intensions Not Confirmed (ns)
Psychological Strain Confirmed
175
5 H3 Emotional
Dissonance
Job Stress Confirmed
Fully
Confirmed
Job Burnout Confirmed
Turnover Intensions Confirmed
Psychological Strain Confirmed
6 H4a Workplace
Bullying
Perceived
Organization Support
Confirmed Fully
Confirmed
7 H4b Workplace
Bullying
Perceived
Co-worker Support
Confirmed Fully
Confirmed
8 H4c Workplace
Bullying
Perceived Supervisor
Support
Confirmed Fully
Confirmed
9 H5 Workplace
Bullying
Emotional
Dissonance
Confirmed Fully
Confirmed
b. Mediating Effect Results (Simple/Parallel and Sequential)
Sr
No
Hyp No IV Mediator/(s) DV Result for
each outcome
Overall
Result
Simple and Parallel mediation Results
1 H6a Workplace
Bullying
Perceived
Organization
Support
Job Stress Not
Confirmed
(ns)
Confirmed
for all
outcomes
except Job
Stress
Burnout Confirmed
Psychological
Strain
Confirmed
176
Turnover
Intensions
Confirmed
2 H6b Workplace
Bullying
Perceived Co-
worker Support
Job Stress Not
Confirmed
(ns)
Fully
Confirmed
for Burnout
and Strain
only
Burnout Confirmed
Psychological
Strain
Confirmed
Turnover
Intensions
Not
Confirmed
(ns)
3 H6c Workplace
Bullying
Perceived
Supervisor
Support
Job Stress Not
Confirmed
(ns)
Fully
Confirmed
for Burnout
and Strain
only
Burnout Confirmed
Psychological
Strain
Confirmed
Turnover
Intensions
Not
Confirmed
(ns)
4 H7 Workplace
Bullying
Emotional
Dissonance
Job Stress Confirmed
Fully
Confirmed
Burnout Confirmed
Psychological
Strain
Confirmed
177
Turnover
Intensions
Confirmed
Sequential Mediation Results
5 H8a Workplace
Bullying
Perceived
Supervisor
Support (M1)
Emotional
Dissonance
(M2)
Job Stress Confirmed
Fully
confirmed
Burnout Confirmed
Psychological
Strain
Confirmed
Turnover
Intensions
Confirmed
6 H8b Workplace
Bullying
Perceived Co-
worker Support
(M1)
Emotional
Dissonance
(M2)
Job Stress Confirmed
Fully
confirmed
Burnout Confirmed
Psychological
Strain
Confirmed
Turnover
Intensions
Confirmed
7 H8c Workplace
Bullying
Perceived
Organization
Support (M1)
Emotional
Dissonance
(M2)
Job Stress Not
Confirmed
(ns)
Fully
Rejected
Burnout Not
Confirmed
(ns)
Psychological
Strain
Not
Confirmed
(ns)
178
Turnover
Intensions
Not
Confirmed
(ns)
c. Moderation Results
Sr
No
Hyp No IV Moderator DV Result for each outcome
1 H9a
Workplace
Bullying
Core Self
Evaluations
Perceived
Organization
Support
Not Confirmed (ns)
2 H9b
Perceived Co-
worker
Support
Not Confirmed (ns)
3 H9c
Perceived
Supervisor
Support
Partially Supported
(opposite effects)
4 H10 Emotional
Dissonance
Partially Supported
(opposite effects)
Chapter Summary
The present chapter highlighted the details of the various data analytical techniques used and the
results of these techniques. Descriptive statistics and correlation analysis were performed. The
study's proposed hypothesis which included the main effects, simple mediation, parallel and
sequential mediation and moderation results were tested through Structural Equation Modeling
179
and Process technique by Hayes. The results of SEM and Process lend support for most of the
direct effects hypothesis except for three support types (POS, PSS and PCS) effects on job stress
and turnover intensions.
The simple mediation results and parallel mediation path of emotional dissonance between
workplace bullying was fully supported. The simple mediation results and parallel path of
perceived organization support between workplace bullying and outcomes was supported for all
outcomes except job stress. The simple mediation/parallel path of perceived co-worker support
and perceived supervisor support between workplace bullying and outcomes was supported for
all outcomes except job stress and turnover intensions.
The sequential mediation results depicted that perceived supervisory and perceived co-worker
support was sequentially related to emotional dissonance and further outcomes whereas
Perceived organization support was unrelated to emotional dissonance. Therefore the results of
the sequential mediation highlight that Perceived supervisor and Emotional dissonance fully
sequentially mediated the relationship between bullying and outcomes. Similarly Perceived co-
worker support and Emotional dissonance fully sequentially mediated the relationship between
workplace bullying and outcomes whereas the results of Perceived organization support were
rejected.
The moderation results showed that the interaction term of workplace bullying and core self
evaluations was significant for two out of four outcomes i.e. Emotional Dissonance and for
Perceived Supervisor Support (moderation is suggested at first link of the research model
therefore outcomes in the case of moderator are Emotional dissonance, Perceived Organization
Support, Perceived Co-worker Support and Perceived Supervisor Support). However, the results
180
of the interaction plots depict that the moderator was significant in the proposed direction
however the relationship between the independent and dependant variables were reversed
suggesting partial support for these two moderation hypotheses. Thus, the results depict support
for most of the main effects, simple mediation, parallel mediation and sequential mediation
according to both SEM and Process methods whereas moderation results were partially
supported.
181
CHAPTER 5
DISCUSSION
Chapter Overview
The current chapter gives the details of the findings of the present study for both SEM and
Process results. Furthermore, it gives the reader information regarding study’s strengths and
limitations. Then, the findings are discussed in the light of theoretical implications it might have
for further theory development and extension. Then the findings of the study are explained in
terms of the benefits and usefulness it might have for mangers in organizations. The findings are
also explained in terms of the relevance of the research model, theory and results in the Pakistani
Context. Lastly, Future research directions are put forth to indicate how the study can be useful
in generating and extending research.
5.1 Major Findings
5.1.1 Overview
The current thesis investigated the impact of workplace bullying as a stressor in creating job
strains through the underlying mechanisms of emotional dissonance and perceived support types.
It also investigated the moderating role of core self evaluation as a personality trait between
workplace bullying and emotional dissonance and perceived support types (i.e. POS, PSS and
PCS). This study utilized temporally segregated data with longitudinal research design collected
182
at three different time periods i.e. time-1, time-2 and time-3 with an appropriate time lag of
approximately 2 to 3 months between each of the three time periods. A sample of complete
useable 334 dataset with match responses from the same respondent was obtained for all the 3
time periods.
5.1.2 Reliability and Validity results summary
Data was verified for errors and missing data was dealt accordingly. Before hypothesis testing
preliminary analysis was done to check instrument reliability and validity. Reliability analysis of
each of the instrument items for each construct revealed excellent reliabilities for all the study
variables. The decision for removing a given number of item/(s) from a particular construct was
done in combination with the reliability and the confirmatory factor analysis results. The items of
a particular variable which demonstrated poor factor loadings and whose removal increased the
overall scale reliability were subsequently deleted from the relevant scale. After checking the
reliability of the measures used in the present study a series of CFAs (confirmatory factor
analysis) were conducted to test each construct's convergent and discriminant validity.
The CFA's conducted included different models comprising of alternate pairing of variables
which were measured within one time period and from the same rater for each time period. The
results of the CFA indicated that for each pairing of variables the multiple factor unconstrained
model demonstrated higher acceptable model fit (see table 2 and 3 for detailed CFA results for
time-2 and time-3 and section 3.8 for CFAs). In addition to the time wise alternate paired
comparisons CFAs, a full measurement model was run as an initial step of Structural Equation
modeling. This full measurement model comprised of CFAs conducted of all the study variables
utilized. The full 9 factor measurement model demonstrated higher model fit statistics signaling
183
adequate model fit to the data. Moreover, each item of the variable displayed factor loadings well
above the minimum level of 0.4 highlighting higher convergent validity. The other tests of
convergent validity were also run for example the results of composite reliability of the
constructs revealed that all the variables had values above the minimum acceptable level of 0.6.
The average variance extracted (AVE) values above 0.45 for all variables showed that all the
study variables demonstrated adequate discriminant and convergent validity.
After establishing construct validity and reliability of the study variables, the researcher moved
on to conduct initial analysis for hypothesis testing. Descriptive statistics were run, Means of the
variables were created, correlational analysis and paired correlation comparisons were conducted
to check the association among the study variables. All the current thesis variables were co-
related significantly in the desired direction. Bivariate Correlation consisting of two-tailed
significance demonstrated that correlation value greater than 0.1 was significant for all the
current thesis variables.
5.1.3 Hypotheses Testing through SEM and Process
Hypothesis testing was done through two different techniques in the present study i.e. the direct
effects, simple, parallel and sequential mediation and moderation hypothesis were tested through
Structural Equation Modeling (SEM) and Preacher and Hayes (2004) process technique using
bootstrapping. These two are not only the latest techniques in mediation and moderation research
but they exhibit superiority on a number of fronts over the existing regression method of testing
mediation and moderation.
184
First SEM is conducted where the direct effects hypothesis are tested one by one in the course of
which direct structural models, after that path models for checking simple/parallel mediation and
lastly sequential mediation are tested. In order to support each hypothesis, the statistics for model
fir are stated which signify model fit to the data, For each hypothesis, the model fit statistics are
reported which indicate fit of the model to the data, the label of parameter estimates consisting of
beta values or standardized regression values running from each of the latent or unobserved
variable is reported along with their significance values which either confirms or rejects the
proposed hypothesis.
Finally R-square values for each path are stated to explain the variation caused in the
intervening/mediating or outcome variables (as defined endogenous variables) because of
specific exogenous or independent variables. A full structural model at the end is run which
consolidates all the proposed hypotheses and presents a complete picture of findings of the full
research model with path estimates and significance along with model fit statistics. After the
confirmation of most of the hypothesis through Structural Equation Modeling then the researcher
tested the proposed hypotheses through another current, robust and relevant technique of
conducting regression through process method developed by Preacher and Hayes (2004) for
testing mediation and moderation and is well suited for testing complex models in quantitative
research.
5.1.4 Direct Effects findings through SEM and Process
Table 15 shows the summary of findings of the suggested study hypothesis. Hypothesis 1
suggested that workplace bullying is related positively to outcomes (i.e. Job stress, burnout,
psychological strain and turnover intensions). The results of SEM and Process lend full support
185
for this hypothesis. H2a states the direct effects of perceptions of organization support on
outcomes. The findings of the Hayes process method and the structural paths revealed that
perceptions of organization support was negatively and significantly associated with all
outcomes except job stress. Thus H2a was supported for all outcomes except job stress. H2b also
deals with the main effects of perceptions of coworker support on outcomes. The findings
highlighted that coworker perceptions of support negatively and significantly related to two out
of the four outcomes i.e. support was found for this hypothesis for burnout and psychological
strain only. H2c proposed that perceptions of supervisory support has a direct negative influence
on job strains. The findings showed that supervisor perceptions of support was negatively and
significantly associated with burnout and psychological strain.
H3 put forth the direct and positive influence of emotional dissonance on work outcomes. The
SEM structural path models and Hayes Process method provided full support for this contention,
as Emotional Dissonance was significantly positively associated to all outcomes proving
hypothesis 3. H4 a, b and c proposed the direct negative effects of workplace bullying on
perceived organization support, coworker perceptions of support and supervisor perceptions of
support correspondingly. The results of SEM and Hayes method also provided full support for
these hypothesis. H5 proposed the direct and positive effect of workplace bullying in creating
emotional dissonance. The SEM and Hayes Process results highlighted that workplace bullying
was significantly positively related to Emotional Dissonance thus rendering full support for
hypothesis 5.
186
5.1.5 Mediation Effects results (Simple, Parallel and Sequential)using Process and SEM
Confirmation of direct effects for almost all of the hypothesis resulted in testing of the simple
and parallel mediation effects hypotheses. H6a proposed the mediating and parallel path of
perceived organization support linking bullying at the workplace and job strains. The results
highlight that POS fully acted as a mediator between work related bullying and burnout,
psychological strain and turnover intensions excluding job stress. Thus, H6a was supported for
all outcomes except job stress. H6b contended that perceived co-worker support acts as a
mediator between workplace bullying and outcomes. The findings reflected that perceived co-
worker support mediated the relationship of workplace bullying on outcomes except job stress
and turnover intensions. Thus, the mediation hypothesis H6b was supported for burnout and
psychological strain only. H6c proposed that the third perceived support type i.e. perceived
supervisor support mediates the relationship of workplace bullying and job outcomes. The
findings of Process and SEM highlight that PSS mediated the relationship of workplace bullying
on all outcomes except for job stress and turnover intensions. H7 stated that emotional
dissonance acts as a mediator in the relationship of workplace bullying on outcomes. Both SEM
and Process method lend full support for this hypothesis as Emotional Dissonance fully mediated
the relationship of workplace bullying on all outcomes.
H8 (a, b and c) proposes that perceived support types (i.e. PSS, PCS and POS) respectively and
Emotional Dissonance sequentially acts as a mediator in the relationship of bullying at the
workplace on job strains. The results depict that PSS, PCS and Emotional Dissonance (H8a and
8b) fully sequentially mediated the relationship of bullying at work on all outcomes. Whereas
Perceived Organization Support was not related to Emotional Dissonance and accordingly did
187
not show effects of sequential mediation between workplace bullying and all job outcomes hence
indicating full rejection of H8c. Thus, 8a and b were fully supported whereas 8c was rejected.
5.1.6 Moderation results through Process method
Moderation effects were checked by using Preacher and Hayes (2004) process method which is
different from the traditional approach of Baron and Kenny (1986) for testing moderation
analysis by means of regression method. It has been corroborated that the Hayes technique
exhibits statistical and procedural advantage above the Baron and Kenny's (1986) conventional
method for performing and checking moderation effects (see Section 4.3.4.1). H9 and H10
pertain to the moderation hypothesis of the study.
Table 15c demonstrated the results of the moderation analysis conducted through the Hayes
Process technique. As depicted in this table H9c and H10 proposed that CSE would moderate the
negative relationship of workplace bullying on perceived supervisor support and positive
relationship of workplace bullying on emotional dissonance such that these relationships would
be stronger for individuals in low core self evaluations respectively. Although the interaction
term of CSE x PSS and CSE x ED was significant however the results of the interaction plots
and subsequent slope tests revealed that workplace bullying and perceived supervisor support
and bullying and emotional dissonance demonstrated opposite directions in case of low core self
evaluations. Thus, H9c and H10 were partially supported whereas H9a and H9b were fully
rejected.
188
On the whole, out of the 10 stated hypothesis, the present thesis confirmed most of the
hypothesis. Complete support was found for 5 out of the 10 hypothesis for all outcomes i.e. H1,
H3, H4 a, b and c, H5 and H7. 3 hypothesis i.e. H2 a, b and c; H6 a, b and c and H8 a, b and c
were supported for all outcomes except job stress and turnover intensions. Finally, the last 2
hypotheses concerning moderation indicated partial support for H9c and H10 whereas H9 a and
b were not confirmed. Thus, support for found for main effects, almost all simple/parallel
mediation and sequential mediation hypotheses whereas moderation results depicted only partial
support.
5.2 Strengths of Study
The current thesis presents several theoretical and methodological advantages. This thesis
focused on studying workplace bullying which is nowadays considered as the most pertinent,
burning and suitable topics not only in management and OB literatures but also in the context of
the global workplace where hostile treatment from superiors, co-workers, customers and sub-
ordinates have become widespread (Einarsen, Raknes & Matthiesen, 1994; Nielsen & Einarsen,
2012; Samnani & Singh, 2012; Tuckey & Neal, 2014).
The present thesis is distinctive as it put forth and then also investigates novel and unexplored
underlying processes of how and why workplace bullying as a stressor leads to strains.
Specifically, it contends that victims of bullying might develop negative perceptions of support
types (i.e. POS, PCS and PSS) which ultimately results in stress, burnout, psychological strain
and turnover intensions.
189
Recent meta-analysis and reviews on workplace bullying suggest weak and in most cases
inconsistent relationships between workplace bullying and outcomes suggesting the possibility of
mediators and moderators existing in this relationship (Nielsen & Einarsen, 2012; Samnani &
Singh, 2012). The present study provides cognitive (POS, PCS and PSS) and affective
mechanisms (Emotional Dissonance) through which workplace bullying shows its disastrous
effects on employees in the form of stress, burnout, psychological strain and turnover intensions.
In line with this, the results also support the full mediation of Perceived Organization Support
between workplace bullying on all outcomes except job stress. Regarding the results of other two
support types full support was found for the mediating influence of perceptions of coworker and
perceptions of supervisory support types between bullying at work and burnout and strain
whereas the other two outcomes of stress and turnover intensions was insignificant. In addition,
full support was found for the results of main, simple mediation and parallel mediation of
emotional dissonance between workplace bullying and outcomes.
Further, two out of the 3 support types i.e. Perceived Co-worker and Perceived Supervisor
support and Emotional dissonance sequentially explain how targets of bullying develop negative
perceptions of supervisor and co-worker support towards them which leads to emotional
dissonance of felt and displayed emotions thus eventually becoming evident in the form of
higher stress, burnout, psychological strain and turnover intensions. Therefore, the present thesis
attempts to offer attractive solutions to resolve queries of 'how' and 'why' workplace bullying as a
stressor creates strains. The current study by proposing and investigating new mechanisms opens
up a new avenue for researchers in the aggression literature and specifically workplace bullying
domain.
190
The current thesis is also unique in the sense as it employs and expand two famous theoretical
approaches in the stress literatures domains that is Conservation of Resources theory (Hobfoll,
1998; 2001) and Cognitive Appraisal theory of Stress and coping (Lazarus & Folkman, 1984) as
overarching theoretical paradigms to justify the theoretical foundations of the suggested research
model links. The solid reasoning employed to develop justifications for hypothesized research
model associations provide confidence and extension of these two theories.
The present study also offers a distinctive edge as it employed research design using longitudinal
and temporally segregated data at three different time points i-e time 1, time 2 and time 3 with a
time interval of approx two months which makes it empirically better than cross-sectional
research examinations of workplace bullying. Nielsen and Einarsen (2012) in their recent meta-
analytical study indicated that till date less than 15 studies exist on workplace bullying all in the
developed countries that have utilized research designs with longitudinally collected data with
one or more time lags. Therefore, the current thesis focuses on an extremely significant and latest
gap in the workplace bullying domain by not only suggesting new underlying processes but also
by using temporally separated data with longitudinal research design at multiple time points with
adequate time gaps for checking the suggested research model relationships.
The present thesis utilized three different time periods with a lag of approx 2 to 3 months in
between for collecting data on the suggested research model. Using even 2 time waves becomes
a tedious and time consuming task, this study employed 3 time periods with a 2 to 3 months time
lag between these three time periods for data collection purposes which makes it empirically
strong. This study employed two of the recent and vigorous techniques for analyzing data
namely Process approach utilizing bootstrapping (Preacher & Hayes, 2004) and Structural
equation modeling (SEM) technique for confirming hypotheses.
191
The results of the study and substantiation for most of the direct paths, simple/parallel mediation
and sequential mediation highlight the significance and methodological superiority of utilizing
these approaches and for analyzing and checking complicated path models. The findings of
confirmatory factor analysis and confirmation for most of the suggested main effects,
simple/parallel and sequential mediation effects signify that the measures employed for checking
the research model variables demonstrate high validity (convergent as well as discriminant) and
reliability and hypothesis are supported expanding the conceptual and methodological field of
investigation in the workplace bullying literatures.
5.3 Study limitations and Future Research Avenues
Although the current research provides novel theoretical insights and methodological vigor, just
like any other research there are weaknesses as well which if attended to by future researchers
can present an opportunity for expanding the workplace bullying domain. First of all, despite the
current study utilized research design with longitudinally collected data at more than one time
points it cannot be classified as a full longitudinal design as all the research model variables were
not tapped at all 3 different time points. The present research can be categorized as study
employing temporally separated longitudinal data collected at three different time periods as the
independent variable of workplace bullying was measured at time-1, the moderator variable of
core self evaluation tapped at time-2 only and the study's mediator variables i.e. perceptions of
support kinds (POS, PCS, PSS) and Emotional Dissonance were tapped at time 2 and time-3 and
the outcome variables along with the mediators were measured at time-3. Scholars in future can
192
gain by employing complete longitudinal design involving more than one time periods where all
the research model variables are tapped at all the different time intervals utilized.
Secondly another constraint of the current research is that even though longitudinal research
design at three different time points with suitable time intervals was employed nevertheless all
the variables were tapped from the same source. However, the kind of variables examined in the
present thesis showed that self-reported measures was the best approach to tap these variables as
previous research highlights that Perceived support types, emotional dissonance, stress burnout,
Turnover intensions, core self evaluations and workplace bullying are more accurately reported
through self reports.
Furthermore, our intension was to determine the victims who experienced bullying at the
workplace therefore employees themselves were in a better position to report whether they have
been exposed to incidences of bullying at the workplace or not. Peers and supervisory reports of
bullying might not be beneficial as they might not be aware of it and even if aware might be
apprehensive to report bullying behaviors inflicted by them onto others. Additionally, the strong
support for the findings of factor analysis for both the full CFA model as well as alternate
models performed within one time period (consisting of adequate factor loadings of the items of
the related constructs, the convergent and discriminant validity results) obviously signal that self
reported measures were not a major issue to the results of the study. Nonetheless, future research
studies need to take account not only a victim perspective but a perpetrator viewpoint of
promoting instances of bullying at the workplace.
Thirdly another drawback of the current study was that I did not find confirmation for the
mediation hypothesis for all the three Perceived types (POS, PCS and PSS) between workplace
193
bullying and job stress. One of the justification for the rejection of meditational hypothesis for
perceived support types between bullying and specifically only for job stress outcome indicates
that workplace bullying reflected strong positive direct effects on job stress and the need for
mediation and moderation arises in the case of weak to non-existent relationships. However,
future researchers should explore these 3 support types between workplace bullying and job
stress particularly to deal with such weakness.
Lastly, although the current thesis found support for most of the direct, simple, parallel and
sequential mediation hypothesis this study did not find the moderating role of Core self
evaluation personality trait between bullying and outcomes. Although the interaction term of
CSE x bullying was significant for Emotional Dissonance and PSS, the resulting interaction plots
depict reverse effects for bullying-emotional dissonance and bullying-perceived support types
lending only partial support for these two and non-confirmation for the other two hypothesis.
Future researchers should investigate the moderating influence of CSE between bullying and
these outcomes. As a final point, potential scholars should replicate the suggested research
model in newer, lesser researched and developing countries context to corroborate the above
results in alike cultures like Pakistan.
5.4 Theoretical Implications
For the past few years, there has been rising attention paid by researchers in understanding and
examining the dark side of employee behaviors particularly workplace bullying (Einarsen et.al.,
2003; Hutchinson, Vickers, Jackson, & Wilkes, 2010; Nielsen & Einarsen, 2012; Samnani &
194
Singh, 2012; Tuckey & Neal, 2014). The current research study contributes to this line of
investigation on workplace bullying and offers new contentions for theory building and
expansion in this domain. This research extends the workplace bullying literature by suggesting
novel dynamics through which workplace bullying as a stressor exhibits its deleterious
consequences on strains. Specifically, this study proposed perceived support types (POS, PCS
and PSS) and emotional dissonance as parallel and sequential mediators between workplace
bullying and outcomes. The above underlying mechanisms have not been examined till date to
the best of the researcher's understanding thus this research is unique in the sense that it is
examining these particular processes.
This research study employs two of the most popular theories in the areas of stress namely
Conservation of resources (COR) (Hobfoll, 1989; 2001; 2002) and Cogntive appraisal theory of
stress and coping (Lazarus & Folkman, 1984) to justify the suggested research model
relationships. The COR theory contends that people try to increase, preserve and guard their
current resources; resources can be of many types i.e. monetary resources, cognitive resources,
support and personality resources.
Hobfoll (1989; 2001; 2002) further argues that resource depletion is more pertinent than resource
expansion and consists of 2 basic assumptions of primary and secondary resource loss. I contend
that primary resource loss occurs when individuals who experience emotional battering in the
form of bullying build negative perceptions of organization, supervisor and co-worker support
towards them which further depletes their resources leading to higher levels of stress, burnout,
strain and turnover intensions.
195
The secondary resource loss happens when individuals who experience bullying try to protect
their existing resources and suppress their negative emotions creating emotional dissonance
which results in a vicious cycle of resource depletion resulting in disastrous consequences like
greater stress, turnover intensions, strain and burnout. Thus, the theoretical foundations of COR
and Cognitive Appraisal theory of stress and coping addresses the main, mediation, parallel and
sequential mediation hypothesis of perceived support types between bullying and outcomes on
one hand and emotional dissonance between workplace bullying and outcomes on the other hand
most of which are supported except for job stress. Hence, the theoretical justifications and the
subsequent data collection and analysis builds confidence for the suggested relationships and
expands the Appraisal theory of stress and Conservation of resources framework.
The current research adds to the perceived support types literatures by proposing workplace
bullying as a significant determinant of perceived support types (POS, PSS and PCS) and further
these support types mediate the relationship of workplace bulling and strains. Past research has
suggested that stressful environments promote incidences of workplace bullying (Einarsen &
Mikkelsen, 2003; Hoel & Cooper, 2000; Vartia, 1996; Zapf et.al., 1996). This study opens a new
avenue in the literature on workplace bullying by asserting that victims experiences of bullying
leads to the development of negative perceptions of not only towards the organization but
towards the supervisor and co-worker's support as well as victims assume that the organization is
not doing to stop such incidences and supervisors and co-workers are also adding fuel to fire.
The consequent results of Hayes Process and SEM confirm these contentions and expand the
COR theory's assumptions as well. Secondly, the present study contributes to the literature on
emotions by proposing emotional dissonance as an outcome for victims exposed to bullying and
further conceptualizes and tests the mediating role of emotional dissonance between bullying and
196
outcomes. The support for emotional dissonance as an outcome of bullying is consistent with the
theoretical dynamics of COR and Appraisal theory of stress and the resulting full support for
emotional dissonance as an outcome of bullying and as a mediator between bullying and
outcomes gives confidence to these proposed dynamics and bridges an important gap by linking
workplace bullying to emotional dissonance literatures.
Future researchers can investigate other emotions as mediating mechanisms between workplace
bullying and outcomes, for example it might be useful to investigate the mediating role of
negative emotions such as anger between workplace bullying and strains. Moreover, on the basis
of the assumptions of COR and Appraisal theory of stress and coping it can be proposed that
future researchers can examine how workplace bullying depletes other individual resources such
as self-identity and self-concept which further creates resource depletion in the form of stress,
burnout, psychological strain and turnover intensions.
5.5 Managerial Implications
This research thesis offers several recommendations for managers working in organizations. The
current study confirmed that employees subjected to instances of bullying at the workplace
develop negative perceptions of organization, supervisor and co-worker support which takes a
toll at employees in the form of higher stress, burnout, strain and turnover intensions. Managers
at the workplace must engage at all levels to curb instances of bullying since it creates negative
employee perceptions of support towards the organization.
197
Managers within organizations should take a number of actions to lessen incidences of
maltreatment in the form of bullying to employees. Managers must be assigned the task of
establishing, clearly conveying and approving system of policy guidelines and rules concerning
interpersonal behavior within organizations. There must be equitable and just procedures
established by means of adequate contribution, involvement and agreement of employees at all
levels in the organization to tackle and curb incidences of workplace bullying. Official liability
procedures should be operational whereby employees despite of their position and status are
answerable and accountable for their misconduct to concerned authorities and departments.
Human resource management units/divisions within corporations should initiate the following
steps to curb workplace bullying (1) Establish official SOPs (standard operating procedures)
signaling a zero tolerance policy towards unpleasant actions such as workplace bullying (2)
Observe firm conformance to these SOPs on a constant basis (3) Development of departments
and divisions whereby victims of such abusive treatment can anonymously register such
instances (4) Preservation of written evidence of such grievances (5) Creation of just and
impartial teams for performing and dealing with disciplinary matters regarding workplace
bullying (6) Periodic follow-up policies to ensure moral and well-timed solutions for both the
bully and the bullied (7) Enforcing punishments in accordance with the kind and degree of
transgression caused by perpetrators of bullying.
Co-worker and witnesses evidencing bullying instances should help victims to report such
incidents as the study hypothesis and findings depict that workplace bullying promotes negative
perceptions of co-worker support as well which is evident in the form of disastrous strains and
quitting of job. Therefore, co-workers should also play their due role in combating bullying
incidents and avoid indulging in ridicule behaviors themselves since the study lends credence to
198
the assertion that targets of bullying also build negative perceptions of co-worker support as
well. Thus co-workers should not only restrain from bullying others themselves, should avoid
being the silent witness and in worse case supporting the perpetrators of bullying against the
victims as the present study lends credence that targets of abuse blame not only the organization
but the supervisors and co-workers in the form of negative perceptions of organization,
supervisor and co-worker support.
The present study also asserted and fully proved that individuals exposed to bullying suppress
their felt negative emotions and display positive emotions to cope with the lost resources creating
emotional dissonance which further exhausts their resources in the form of greater stress,
burnout, turnover intensions and strain. Managers and organizations should arrange stress
management and emotional management workshops to help employees vent their frustrations
particularly for employees who have been subjects of emotional abuse and verbal mistreatment.
These employees would let out their suppressed negative emotions as this study explains that
bullied employees feel emotional dissonance which further uses up their emotional resources
creating strains and eventually quitting the organization.
5.6 Contextual Implications
Quantative research studies in the area of workplace bullying conducted in a number of countries
across the globe such US, UK, Germany, Canada, Argentina, Nordic Countries, Australia,
Turkey, India, Hong Kong, Singapore etc indicates the nature and prevalence of bullying in
different occupational contexts and within different cultural contexts (Einarsen & Skogstad,
199
1996; Hoel, Cooper, & Faragher, 2001; Hutchinson & Eveline, 2010; Jimenez, Munoz, Gamarra,
& Herrer, 2007; Keashly & Jagatic, 2000; Power, et al., 2013).
Nonetheless, majority of research on workplace bullying signify that few studies are present in
this area in the developing countries context such as Pakistan. The current study not only attends
to an extremely pertinent and global burning topic but also advocates novel processes of how
workplace bullying as a stressor manifests itself in employees in the form of strains in a
developing country context like Pakistan. The present thesis has also a distinctive edge as it
employs temporally segregated longitudinal research design with three time wave data to
confirm novel theoretical dynamics in a newer and lesser studied context like Pakistan.
With respect to the insights into national culture, Hofstede (1983) more than 40 years ago carried
an extensive survey of employees working in the subsidiaries of IBM corporation in more than
50 different countries and came with a five dimensional classification of national cultures. Since
that time the acceptability and usefulness of these cultural dimensions have spawned a lot of
research and each country differs on scoring of these cultural dimensions (Bond, 1988; Hofstede,
1983; Triandis, 1982; Trompenaars, 1983; Schwartz, 1994; Smith, Dugan & Trompenaars,
1996). Although it was not the objective of the present study to measure national cultural
dimensions with respect to bullying and is beyond the scope of the study, however the present
study's research model findings can only be explained and applied in the current cultural context
by using the Hofstede's classification of national culture.
Furthermore, the applicability and relevance of findings on any area in the management and OB
domains is incomplete and presents an imperfect picture of the reality unless and until it is
discussed with respect to the connotations that study's findings might have on a particular
200
culture. This study proposed, tested and unfolded an important and new mechanism of Emotional
Dissonance as an outcome of workplace bullying and also Emotional dissonance as a mediator
between workplace bullying and outcomes. The main effects, simple mediation, parallel and
sequential mediation effects were fully proved for Emotional Dissonance for all the outcomes in
the present study. One of the major connotations for full support of this hypothesis in a cultural
context such as Pakistan is that according to Hofstede (1983, 1991), Pakistan has a large power
distance which means that individuals living in such societies not only accept but except that
power such as authority, status etc to unequally distributed at all levels in the society.
Employees in such cultures when exposed to workplace bullying since perceive a high power
gap between them and others particularly the authorities therefore these bullied employees
suppress their negative emotions creating emotional dissonance which further becomes evident
in the form of higher stress, burnout, psychological strain and greater turnover intensions.
Another reason for the support of emotional dissonance as a mediator between workplace
bullying and outcomes is that Pakistan also categorizes as a highly collectivist culture.
Therefore, the application of findings of the above hypothesis in the Pakistani cultural context
would imply that when victims are subjected to ridicule behaviors in the form of bullying from
co-workers the targets of such incidences are supposed to tolerate such humiliating behaviors in
the name of personal friendships and loyalty. Thus, individuals experiencing bullying from co-
workers/peers generate emotional dissonance due to suppression of emotions in order to maintain
in-group affiliations and personal friendships which could take a toll on the part of such
employees resulting in higher incidences of stress, burnout, turnover intensions and
psychological strain.
201
The current thesis also advocated, tested and proved that bullying leads to the development of
negative perceptions of organization, supervisor and co-worker support which further affects
outcomes. This thesis found confirmation for the mediating effects of the three support types
between workplace bullying and all outcomes except job stress and turnover intensions. Again
applying the findings of this hypothesis in the Pakistani cultural context would mean that due to
having collectivist orientations when employees are subjected to bullying behaviors since they
expect from in-groups these victims of abuse might develop negative perceptions of co-worker
support. In addition, Pakistan also scores strong on uncertainty avoidance and short -term
orientation dimensions which means that employees working in the organizations of such
cultures since perceive a highly limited view of their organizations and supervisors, when such
individuals are exposed to bullying at work might immediately develop negative perceptions not
only towards the organizational representatives but also towards the organization itself. Such
targets of bullying due to having a risk averse and narrow perspective of others might blame the
organization that it does not support and care for them and is not making any effort to stop such
incidences of bullying which might particularly result in higher levels of burnout and
psychological strain as well.
Regarding the sequential mediation hypothesis, this study hypothesized that three support types
would lead to emotional dissonance which would create stress, strain, turnover intensions and
burnout. The results of the present study demonstrate that perceived co-worker and perceived
support fully sequentially mediate the relationship between workplace bullying and outcomes
whereas POS although showed simple and parallel mediation, POS did not sequential mediate
this relationship. The connotation for this finding in the present culture would suggest that in a
collectivist culture such as Pakistan where people work in strong in-groups and exhibit greater
202
loyalty towards others when exposed to maltreatment in the form of bullying would not only
develop negative perceptions towards the supervisor but also towards co-workers as well which
would induce emotional dissonance resulting in job strains.
Individuals in such cultures as a result of exposure to bullying although might develop negative
perceptions towards the organization as well but this would not induce emotional dissonance and
negative POS might directly create stress, burnout, strain and turnover intentions. Thus, POS did
not sequentially mediate with ED between bullying and outcomes whereas PSS and PCS fully
sequentially mediate along with emotional dissonance between bullying and outcomes in the
current cultural context.
5.7 Conclusion
Workplace bullying has surfaced as one of the most important and burning topics of research in
OB and management areas for the past many years. This controversial subject has not only
garnered rising interest but is also drawing practitioner's attention as well (Einarsen &
Mikkelsen, 2003; Nielsen & Einarsen, 2012; Samnani & Singh, 2012; Tuckey & Neall, 2014).
Workplaces today are marred with instances of bullying and keeping in view the kind, degree
and occurrence of this phenomenon in a variety of work environments, this study attempts to fill
a significant gap in the workplace bullying literatures by proposing bullying as a stressor which
creates strains through unique mechanisms.
Although the workplace bullying as a stressor creating strains is well-established (Einarsen &
Mikkelsen, 2003; Zapf, Knorz, & Kulla, 1996b) but the processes as a result of which bullying
203
shows its deleterious effects on strains is under-researched (Nielsen & Einarsen, 2012; Samnani
& Singh, 2012; Tuckey & Neall, 2014). This study opens up a new avenue in the bullying
literatures by advocating perceptions of support types (POS, PCS and PSS) as cognitive
processes through which bullying influences work related outcomes. Another key contribution of
the present thesis is that it asserts emotional dissonance as one of the affective mechanisms
through which workplace bullying manifests itself in the victims in the form of greater stress,
burnout, strain and turnover intensions. The current study investigated both the parallel as well
sequential mediation effects of Perceived Support types (POS, PCS and PSS) and Emotional
Dissonance between workplace bullying and outcomes.
This study employed the basis of two of the most renowned theories namely Cognitive appraisal
theory of stress (Lazarus & Folkman, 1984) and Conservation of resources (Hobfoll, 1988; 1989;
2002) to justify the suggested research model relationships. A research design comprising
longitudinal and temporally segregated data was used to collect data at three different time point
i-e time 1, time 2 and time 3 with an adequate time interval of approx 3 months between each.
Different data analysis techniques were employed such as descriptive statistics, validity and
reliability tests, correlational tests, Confirmatory factor analysis. In addition, two of the most
vigorous statistical approaches of Process by Hayes (2004) and SEM were used to check and
verify hypothesis. The Hayes regression technique and structural models of SEM substantiated
majority of the direct, simple/parallel mediation and sequential mediation hypotheses. The
results of the moderation analysis indicated only partial support. Hence, in general the current
thesis offers novel mechanisms of how bullying at the workplace negatively influences
employee's job outcomes and contributes to not only the workplace bullying literature but also
204
expands the methodological base by checking these new mechanisms in an unexplored and new
context of Pakistan.