Transcript

This article was downloaded by: [University of Chicago Library]On: 19 October 2014, At: 04:04Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954Registered office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH,UK

European Journal of Work andOrganizational PsychologyPublication details, including instructions for authorsand subscription information:http://www.tandfonline.com/loi/pewo20

Role stressors and exposure toworkplace bullying: Causes orconsequences of what and why?Lars Johan Hauge a , Anders Skogstad b & StåleEinarsen ba National Institute of Occupational Health, Oslo,Norway, and University of Bergen , Bergen, Norwayb University of Bergen , Bergen, NorwayPublished online: 19 Aug 2010.

To cite this article: Lars Johan Hauge , Anders Skogstad & Ståle Einarsen (2011) Rolestressors and exposure to workplace bullying: Causes or consequences of what andwhy?, European Journal of Work and Organizational Psychology, 20:5, 610-630, DOI:10.1080/1359432X.2010.482264

To link to this article: http://dx.doi.org/10.1080/1359432X.2010.482264

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all theinformation (the “Content”) contained in the publications on our platform.However, Taylor & Francis, our agents, and our licensors make norepresentations or warranties whatsoever as to the accuracy, completeness, orsuitability for any purpose of the Content. Any opinions and views expressedin this publication are the opinions and views of the authors, and are not theviews of or endorsed by Taylor & Francis. The accuracy of the Content shouldnot be relied upon and should be independently verified with primary sourcesof information. Taylor and Francis shall not be liable for any losses, actions,claims, proceedings, demands, costs, expenses, damages, and other liabilitieswhatsoever or howsoever caused arising directly or indirectly in connectionwith, in relation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes.Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expresslyforbidden. Terms & Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

Role stressors and exposure to workplace bullying:

Causes or consequences of what and why?

Lars Johan HaugeNational Institute of Occupational Health, Oslo, Norway, and University of

Bergen, Bergen, Norway

Anders Skogstad and Stale EinarsenUniversity of Bergen, Bergen, Norway

Stressful working environments are commonly assumed to create conditionsthat can lead to bullying. Although environmental factors may relate to bothexposure to and perpetration of workplace bullying, empirical knowledge isstill limited regarding how environmental factors and bullying is causallylinked. In line with interpretations of previous cross-sectional findings, thepresent study investigated an individual-level target-oriented model of rolestressors as antecedents of exposure to workplace bullying, applying a two-wave longitudinal sample of the Norwegian working population. Results fromstructural regression analyses failed to identify any significant relationshipsbetween the role stressors and subsequent exposure to workplace bullying.However, prior exposure to workplace bullying accounted for subsequentvariation in role ambiguity, role conflict, and role overload alike. The findingsof the present study question conclusions regarding causality made in previouscross-sectional research on work-related factors as antecedents of exposure toworkplace bullying. We argue that future research may benefit from adoptingmultilevel and actor-oriented perspectives in order to more fully understandhow causal relationships between environmental factors and workplacebullying unfold.

Correspondence should be addressed to Lars Johan Hauge, National Institute of

Occupational Health, Gydas vei 8, N-0033 Oslo, Norway. E-mail: [email protected]

This study is a collaborative project between the University of Bergen and Statistics

Norway. The project was made possible by a grant from the Research Council of Norway and

by joint grants from two Norwegian employer associations (the Confederation of Norwegian

Enterprise and the Norwegian Association of Local and Regional Authorities), the Norwegian

government (the National Insurance Administration), and its FARVE programme. Thanks to

Bengt O. Lagerstrøm and Maria Høstmark of Statistics Norway, and Stig B. Matthiesen and

Morten B. Nielsen of the Faculty of Psychology, University of Bergen, for their contribution to

the data collection. Thanks also to Professor Evert van de Vliert for his valuable feedback.

EUROPEAN JOURNAL OF WORK AND

ORGANIZATIONAL PSYCHOLOGY

2011, 20 (5), 610–630

� 2011 Psychology Press, an imprint of the Taylor & Francis Group, an Informa business

http://www.psypress.com/ejwop http://dx.doi.org/10.1080/1359432X.2010.482264

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

Keywords: Bullying; Cross-lagged; Harassment; Longitudinal; Role stress.

Workplace bullying has repeatedly been shown to have detrimentalconsequences for affected individuals and to have wide-ranging negativeconsequences for organizations at large (Aquino & Thau, 2009; Hoel, Zapf, &Cooper, 2002). Thus, gaining a thorough understanding of why and howbullying develops in workplaces is of utmost importance in terms of preventingand managing the problem. Several theoretical explanations can be applied toaccount for the development of workplace bullying, but prevailing explana-tions all emphasize the importance of organizational and work-related factorsas underlying in processes leading up to bullying (cf. Einarsen, Hoel, Zapf, &Cooper, 2003; Leymann, 1996; Zapf, 1999). However, most studies testingsuch relationships have applied cross-sectional data, and firm evidence istherefore lacking regarding causal relationships between the variables ofinterest. To make up for this shortcoming, the aim of the present study is totest assumed causal relationships in line with interpretations of previous cross-sectional findings and a target-oriented model as proposed by Bowling andBeehr (2006), regarding role stressors as antecedents of subsequent exposure toworkplace bullying. Correspondingly, the possible reverse relationshipbetween prior exposure to workplace bullying and subsequent experiencedrole stress will also be tested.

Workplace bullying consists of repeated and prolonged exposure tovarious forms of mainly psychological mistreatment directed at a target whois typically teased, badgered, and insulted, and who perceives him- or herselfas lacking the required resources to retaliate in kind (Einarsen et al., 2003).This implies an actual or perceived power imbalance between the partiesinvolved, whereby an individual ends up in an inferior position and becomesthe target of systematic negative social acts (Brodsky, 1976). Althoughexposure to such mistreatment appears under many different labels such asharassment (e.g., Brodsky, 1976), mobbing (e.g., Leymann, 1996), emotionalabuse (e.g., Keashly & Harvey, 2005), and victimization (e.g., Aquino &Thau, 2009), they all seem to refer to the same overall phenomenon, labelledas workplace bullying in this study, namely systematic and prolongedmistreatment of an organizational member, which over time may result insevere consequences for the affected individual (Einarsen et al., 2003).Whether carried out deliberately or unintentionally, workplace bullying canbe understood as a gradually evolving process in which the target in the earlyphases of the process is being exposed to subtle and often disguised forms ofmistreatment, whereas, later on, more direct and openly aggressive behaviourmay occur (Einarsen, 1999; Zapf & Gross, 2001). Workplace bullying cantake the form of direct acts, such as verbal abuse, accusations, and publichumiliation, but it can also be of a more subtle and indirect nature in the

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 611

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

form of gossiping, rumour spreading, and social exclusion (Einarsen et al.,2003). When frequently and persistently directed at the same individual,however, even such subtle and indirect behaviour can be experienced as anextreme source of social stress at work (Zapf, 1999). Nevertheless, therepetitive and unwanted nature of the behaviour involved is essential to theconcept of workplace bullying. Exposure to such behaviour is resented bythe target and is associated with distress and humiliation in both the affectedtarget and observers of such behaviour, implying that workplace bullyingmay have more far-reaching direct and indirect consequences for bothindividuals and organizations in addition to those being directly exposed(Hauge, Skogstad, & Einarsen, 2007; Vartia, 2001).

The reasons why bullying develops can be complex and interwoven andmay be related to both characteristics of the targeted individual and tocharacteristics of the perpetrator (Baillien, Neyens, de Witte, & de Cuyper,2009; Zapf, 1999). Nevertheless, a common explanation in most theoreticalframeworks is the presence of underlying problematic work environmentconditions that create a fertile ground for bullying to occur (Bowling &Beehr, 2006; Keashly & Harvey, 2005). In this sense, the work environmenthypothesis, which states that stressful and poorly organized workingenvironments may give rise to conditions resulting in bullying, hastraditionally been the favoured causal model when aiming to explain theoccurrence of workplace bullying (Agervold & Mikkelsen, 2004; Leymann,1996). According to this overall situational view, workplace bullying isprimarily caused by factors related to deficiencies in work design and toleadership behaviour within organizations (Leymann, 1996). Especiallywhen combined with a management that avoids or neglects to intervene inand manage stressful situations, unfavourable and problematic environ-mental conditions may cause increased tension and frustration in the workgroup, thereby triggering the bullying process (Zapf & Gross, 2001).

In the workplace bullying literature, work-related factors are commonlyargued to relate to bullying in accordance with two main theoreticalframeworks, namely explanations in line with social interactionist theoryand the frustration-aggression hypothesis (cf. Baillien et al., 2009), where thefirst of these is the main focus of this study. In line with social interactionisttheory (e.g., Neuman & Baron, 2003; Tedeschi & Felson, 1994), individualsmay have a tendency to become targets of bullying because stressors in theirjobs generate affective and behavioural reactions in them that mayencourage others to engage in aggressive behaviour towards them. On theother hand, explanations in accordance with the frustration-aggressionhypothesis (e.g., Berkowitz, 1989; Spector & Fox, 2005) hold that thepresence of environmental stressors may give rise to aggressive behaviour bygenerating negative affect in individuals, thus encouraging perpetrators toengage in bullying of other organizational members. However, who will

612 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

end up as a target of bullying due to perpetrators’ aggressiveness withinsuch stressful working environments can be highly arbitrary (Leymann,1992).

In line with these frameworks, Bowling and Beehr (2006) proposed atarget-oriented model to explain the occurrence of workplace bullying at anindividual level of analysis, focusing on the influence of role stressors asantecedents of bullying. Role stress is often considered to consist of threedistinct but related concepts. By role ambiguity is meant uncertainty aboutwhich actions to take in order to fulfil the expectations of one’s work role,while role conflict arises when the different expectations and demands ofone’s work role are incompatible (Beehr, 1995). The somewhat lessfrequently studied concept of role overload was initially viewed as a subtypeof role conflict (cf. Kahn, Wolfe, Quinn, Snoek, & Rosenthal, 1964), but ithas subsequently been established as a distinct but related concept. Roleoverload can be conceptualized as the extent to which the time andresources available prove inadequate in terms of fulfilling one’s role andmeeting the expectations and obligations associated with it (Beehr, 1995).Unlike task-related stressors such as work pacing and skill under-utilization, role stressors may have something in common with workplacebullying in that they are both inherently social in nature (Bowling &Beehr, 2006). Role stressors represent expectations and demands fromother people at work, whereas workplace bullying consists of negativebehaviour from the same set of people, both representing social stressorswith likely negative consequences for individual well-being (Beehr, 1995;Einarsen, 2000). Role stressors are among the most frequently studiedstressors in occupational psychology in general, and their associationswith individual health and well-being are well documented (e.g., Ortqvist& Wincent, 2006).

According to Bowling and Beehr’s (2006) model, being exposed to rolestressors could produce negative emotional and behavioural responses inindividuals that can encourage victimization. In addition, targets’ experi-ences of a high level of role stress might indicate the presence of ambientstressors experienced by most individuals in the target’s work environment.Thus, the presence of role stressors may cause stressed individuals to engagein aggressive behaviour towards others in their work environment. AlthoughBowling and Beehr apply explanations in line with both a socialinteractionist perspective (e.g., Tedeschi & Felson, 1994) and enactment ofaggression (e.g., Spector & Fox, 2005) to account for how role stressors arerelated to workplace bullying, their meta-analysis undertaken to investigatethe usefulness of the model concerns cross-sectional findings of relationshipsbetween individual perceived role stress and individual exposure toworkplace bullying, findings necessarily in line with target-orientedexplanations such as social interactionist theory.

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 613

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

With the few exceptions focusing on perpetrators of bullying (e.g., DeCuyper, Baillien, & De Witte, 2009; Hauge, Skogstad, & Einarsen, 2009),research has so far mainly focused on the experiences of targets of bullying.In this respect, investigations into environmental causes of workplacebullying have traditionally followed in the footsteps of job-stress research inregarding exposure to workplace bullying as a response to perceivedstressful working conditions (cf. Hoel et al., 2002). As such, several cross-sectional studies have provided evidence for consistent relationships betweenrole stressors and exposure to workplace bullying. For instance, Einarsen,Raknes, and Matthiesen (1994) found role conflict (r¼ .26) to correlaterelatively strongly with exposure to workplace bullying, while a weakercorrelation was found for role ambiguity (r¼ .11). The same pattern ofrelationships was also identified by Notelaers, De Witte, and Einarsen(2010), with role conflict (r¼ .43) showing a somewhat stronger correlationthan did role ambiguity (r¼ .38). In their meta-analysis, Bowling and Beehr(2006) reported that both role conflict (r¼ .44) and role ambiguity (r¼ .30)correlated with workplace harassment. In addition, they identified roleoverload (r¼ .28) as showing a corresponding relationship.

Although most theoretical frameworks on the development of workplacebullying assume organizational and work-related factors (e.g., role stressors) toact as antecedents of bullying and subsequent strain (cf. Bowling & Beehr,2006; Keashly & Harvey, 2005; Leymann, 1996), empirical evidence has notyet been presented for such causal relationships. Seen from a stress-theoreticalperspective on exposure to workplace bullying (e.g., Hoel et al., 2002), rolestress can contribute substantially to the development of bullying due to rolestressors’ effects on targets’ behaviour. However, it may be equally plausiblethat prior exposure to bullying is related to subsequently experienced rolestress in targets (cf. Zapf, 1999). Although cross-sectional designs provide nobasis for making such causal inferences, studies have indeed shown targets ofbullying to generally report significantly higher levels of both role ambiguityand role conflict than do nontargets (e.g., Agervold &Mikkelsen, 2004; Haugeet al., 2007; Jennifer, Cowie, & Ananiadou, 2003; Matthiesen & Einarsen,2007). It is possible that being exposed to aggressive behaviour at work willresult in more general negative attitudes towards one’s work environment, andthat this pattern of relationships thus can be interpreted as a response tostressful working conditions (Agervold, 2009). Following such a line ofreasoning, the experienced role stress due to bullying may make targets evenmore vulnerable of persistently being exposed to bullying in a cyclic ongoingrelationship (cf. Kahn et al., 1964).

However, it is also possible that experienced role stress is a directconsequence of the bullying exposure. In accordance with role theory, whenexpectations and demands directed at the individual are conflicting,ambiguous, or overloading, the individual is expected to experience role

614 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

stress (cf. Beehr, 1995; Kahn et al., 1964). In the same way, behaviourassociated with workplace bullying can also have a direct effect on rolestress. For instance, withholding information or isolating someone socially willundoubtedly complicate the work situation for the affected individual andprobably lead to role ambiguity, and, likewise, being exposed to overwhelmingdemands on top of other already assigned work tasks will probably be relatedto subsequent perceptions of role overload. Role pressure is exerted on anindividual in an attempt to influence or change the behaviour of the individualin a given situation through expectations or demands that can be more or lessefficiently and fairly communicated to the individual in a series of cyclic roleepisodes (Kahn et al., 1964). Thus, pressure in the form of bullying behaviourcan repeatedly be inflicted on the individual in order to make him or hercomply with expectations or demands from members of the role set. Ifperceived as legitimate and manageable, such pressure can be seen as a featureof ordinary organizational interaction, whereas, if perceived as illegitimate orotherwise resented by the recipient, the same pressure may be perceived asbullying (Brodsky, 1976). If the initial pressure towards change in the focalperson’s behaviour does not yield the desired effect, role senders may inducenegative sanctions towards the individual through means of fear andintimidation or threats of expulsion from the social work group (Katz &Kahn, 1978). Thus, role pressure may over time turn into pressure in the formof bullying, further increasing the experienced role stress (cf. Kahn et al.,1964). Some members of the organization may even view bullying asfunctional and legitimate in terms of achieving acceptable productivity andmeeting performance standards (cf. Brodsky, 1976; Ferris, Zinko, Brouer,Buckley, & Harvey, 2007).

In accordance with the theoretical reasoning just outlined, the present studywill test causal relationships between role stressors and exposure to workplacebullying at an individual level of analysis, employing a longitudinal two-wavesample of the Norwegian working population. In line with a socialinteractionist perspective (e.g., Neuman & Baron, 2003) and Bowling andBeehr’s (2006) target-oriented model, we expect experienced role stress to berelated to subsequent exposure to workplace bullying. Moreover, in line withrole theory (cf. Kahn et al., 1964), we expect exposure to workplace bullying tobe related to subsequent experienced role stress.

METHOD

Sample characteristics

The present two-wave study is based on data collected through anonymousself-reporting questionnaires during the spring of 2005 and the spring of2007. A total of 4500 employees, representative of the Norwegian

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 615

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

workforce, were drawn from the Norwegian Central Employee Register byStatistics Norway (SSB). At Time 1 a total of 2539 questionnaires werereturned, yielding a first-wave response rate of 56.4%. Of these respondents,1775 completed questionnaires also at Time 2, yielding a second-waveresponse rate of 70%. Respondents were asked to report their job status atboth measurement points, and only respondents currently employed in full-or part-time positions were included in the analyses (e.g., unemployedemployees or employees retired from work at Time 2 completed only thedemographics section). As the present study aims to investigate relationshipsbetween role stressors and exposure to workplace bullying over time andunder comparable environmental conditions, employees who had changedtheir place of employment from measurement wave 1 to wave 2 were alsoexcluded from the sample. The selection criteria left a cross-sectional sampleof 2242 employees and a longitudinal sample of 1207 employees. The two-wave sample consisted of 52.6% women and 47.4% men. The mean age was46 years, and 21% reported to occupy a supervisory position. With regard torepresentativeness (Table 1), the longitudinal sample consisted of signifi-cantly more women as compared to the representative sample from which itwas drawn, w2¼ 11.21, df¼ 1, p5 .01. The employees constituting thelongitudinal sample were also significantly older, w2¼ 109.90, df¼ 4,p5 .01, with especially employees in younger age categories tending to beunderrepresented (cf. Høstmark & Lagerstrøm, 2006).

Measures

Exposure to workplace bullying was measured by a nine-item short versionof the Negative Acts Questionnaire (Einarsen, Hoel, & Notelaers, 2009;

TABLE 1Representativeness by gender and age

Drawn sample

(n¼ 4500)

2-wave sample

(n¼ 1207)

n % n %

Gender

Women 2121 47.1 635 52.6

Men 2379 52.9 572 47.4

Age

Below 30 years 692 15.4 80 6.6

30–39 years 1196 26.6 266 22.0

40–49 years 1206 26.8 356 29.5

50–59 years 1033 23.0 410 34.0

60 years or more 373 8.3 95 7.9

616 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

Notelaers & Einarsen, 2008). The nine items describe different kinds ofbehaviour that can be perceived as bullying if they occur on a regular basis.The short version consists of three items each concerning social isolation,work-related bullying, and person-related bullying, respectively. All itemsare formulated in behavioural terms with no reference to the term bullying.For each item, respondents were asked to rate how often they had beenexposed to that specific behaviour in their present workplace during theprevious 6 months (e.g., ‘‘Been ignored or excluded from work groupactivities’’). The five response categories applied were ‘‘never’’, ‘‘now andthen’’, ‘‘monthly’’, ‘‘every week’’, and ‘‘daily’’.

Role ambiguity and role conflict were measured using the scales of Rizzo,House, and Lirtzman (1970). Role ambiguity consisted of six items,formulated as role clarity and reverse scored for the analyses, measuringthe existence of clarity of behavioural requirements at work (e.g., ‘‘I knowexactly what my responsibilities are’’). Role conflict consisted of eight items(e.g., ‘‘I receive incompatible requests from two or more people’’) measuringincongruence or incompatibility in the requirements for one’s work role.Role overload was measured by the three-item scale of Beehr, Walsh, andTaber (1976) concerning overloading role expectations in the time available(e.g., ‘‘It often seems like I have too much work for one person to do’’). Allrole stressor scales were measured with seven response categories rangingfrom ‘‘very false’’ to ‘‘very true’’.

Statistical analyses

Statistical analyses were conducted using structural equation modelling inLISREL 8.80 (Joreskog & Sorbom, 1996). Covariance and asymptoticvariance–covariance matrices were analysed to correct for nonnormalityin data by means of robust maximum likelihood estimation. Prior toanalysing longitudinal relationships, attrition analyses were conducted,comparing dropouts from the sample with respondents who responded atboth measurement points, in order to investigate whether disappearancefrom the sample is an outcome of a causal dynamic that is different fromthe remaining respondents. A two-step approach was followed for thelongitudinal analyses (cf. Anderson & Gerbing, 1988; Joreskog, 1993).First, a measurement model was tested for invariance of factor loadingsand indicator intercepts over time (cf. Chan, 2002), and, second, cross-lagged structural equation models were estimated. Longitudinal modelswere estimated without error-autocorrelations of observed indicators.Missing data was handled by the multiple imputation option (MI) inPRELIS 2.80 (Joreskog & Sorbom, 1999) by imputing 10 datasets forsubsequent simultaneous estimation (see, e.g., Graham, 2009, for adiscussion on the use of MI). To assess overall model fit, several

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 617

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

commonly used fit indices were applied (cf. Bentler, 1990; Joreskog, 1993;Steiger, 1990): the chi-square statistic (w2), the root mean square error ofapproximation (RMSEA), the comparative fit index (CFI), and thenonnormed fit index (NNFI). Competing models were compared using ascaled chi-square difference test (see, e.g., Brown, 2006). All coefficientspresented in the following are completely standardized regression paths orcorrelation coefficients.

RESULTS

Cross-sectional analyses

Comparing dropouts from the sample with the remaining respondentsshowed no significant difference with regard to hierarchical position,w2¼ 1.58, df¼ 1, p4 .05, while the dropout group consisted of significantlymore men (52.0%) than the remaining respondents (47.4%), w2¼ 4.52,df¼ 1, p5 .05. The dropout group was also significantly younger(M¼ 41.63, SD¼ 12.08) than the remaining respondents (M¼ 46.01,SD¼ 10.12), t¼ 9.35, df¼ 2240, p5 .01. Systematic attrition for the latentvariables was tested by estimating a MIMIC model regressing a dummyvariable reflecting dropout from the sample on the measurement instru-ments (cf. Brown, 2006). The model showed acceptable fit to data, w2¼ 2062,df¼ 315, RMSEA¼ .050, CFI¼ .965, NNFI¼ .960, and showed that thedropout group did not differ significantly in terms of role ambiguity, roleconflict, or role overload as compared to the remaining respondents.However, the dropout group reported somewhat more exposure toworkplace bullying, estimate¼ .045, SE¼ .016, t¼ 2.81, than did theremaining respondents. Although significantly different, the differencebetween the groups corresponds to a Cohen’s d of only 0.12, an effect sizeconsidered to be very low (Cohen, 1992). Thus, the two groups are quitecomparable in terms of the latent variables, and it can be concluded that thelongitudinal sample does not deviate to any great extent from the cross-sectional sample.

Longitudinal analyses

Prior to longitudinal covariance structure modelling, descriptive statisticsand Cronbach’s alpha coefficients were computed. As can be seen inTable 2, all measurement instruments had alpha coefficients equal to orhigher than .73.

The longitudinal analyses were done in two steps by first estimatingmeasurement models and testing for longitudinal measurement invariance,before proceeding to analyse longitudinal structural relationships between

618 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

the latent variables. Establishing longitudinal measurement equivalence is afundamental aspect of evaluating temporal change in a construct, as itdetermines whether observed change is due to true change or whether thechange is due to changes in the construct itself over time (Brown, 2006;Chan, 2002). To establish the factor structure and to obtain correlationscorrected for random measurement error, a measurement model (M1) of themeasurement instruments at Time 1 and Time 2 was fitted to data. Allcorrelations were significant and test–retest correlations were found in therange between .66 and .78. The measurement model with equal factorloadings over time (M2) did not show a deterioration of fit to data,w2scaled diff ð22Þ ¼ 3:84ns, as compared to Model 1 with no constraints onparameters (Table 3). Constraining the indicator intercepts to equalityover time (M3) also did not show a deterioration of fit to data,w2scaled diff ð22Þ ¼ 32:88ns. The model testing thus indicates longitudinalmeasurement equivalence for exposure to workplace bullying and the rolestressors over time, and thus allows testing for longitudinal structuralrelationships between the factors to proceed.

As shown in Figure 1 (observed indicators of latent factors not depictedfor clarity), autoregressive paths and covariances reflect a stability modelthat is considered to be an important null hypothesis model to be rejectedbefore other causal relationships between the latent variables can be takenseriously (cf. Hertzog & Nesselroade, 1987). The relationships in thestability model are included in all the tested models outlined later. Thelongitudinal structural regression analyses were done in three steps. First, astability model was estimated. Second, a model estimating structural pathsfrom the role stressors at Time 1 to workplace bullying at Time 2 wascompared to the stability model. Finally, a model estimating structuralpaths from workplace bullying at Time 1 to the role stressors at Time 2 was

TABLE 2Descriptive statistics and correlations

Mean SD 1 2 3 4 5 6 7 8

1. Workplace bullying (T1) 1.20 0.29 (.80)

2. Role ambiguity (T1) 2.13 0.87 .29 (.85)

3. Role conflict (T1) 3.07 1.24 .49 .42 (.82)

4. Role overload (T1) 3.64 1.51 .31 .32 .73 (.73)

5. Workplace bullying (T2) 1.19 0.27 .78 .23 .40 .25 (.79)

6. Role ambiguity (T2) 2.16 0.90 .29 .66 .37 .28 .36 (.86)

7. Role conflict (T2) 3.04 1.27 .44 .33 .72 .51 .50 .49 (.84)

8. Role overload (T2) 3.56 1.51 .27 .22 .54 .66 .30 .32 .74 (.75)

Cronbach’s alpha coefficients are presented on the diagonal in parentheses. Correlations are

significant at the p5 .01 level (t+ 2.58); T1¼Time 1; T2¼Time 2.

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 619

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

TA

BL

E3

Str

uct

ura

lre

gre

ssio

ns

be

twe

en

wo

rkp

lace

bu

lly

ing

an

dro

lest

ress

ors

w2df

SCF

Dw2

Ddf

RMSEA

90%

CI

Pclose

fit

CFI

NNFI

Longitudinalmeasurementmodels

M1:Noconstrained

parameters

5072.81

1246

1.207

––

.050

.049–.052

.297

.958

.955

M2:Equalfactorloadings

5078.11

1268

1.210

(M2–M1)3.84ns

22

.050

.048–.051

.536

.958

.956

M3:Equalindicatorintercepts

5110.17

1290

1.206

(M3–M2)32.88ns

22

.050

.048–.051

.694

.958

.957

Longitudinalstructuralmodels

M1:Stabilitymodel

5133.79

1302

1.208

––

.049

.048–.051

.754

.958

.957

M2a:RS(T1)!

WB(T2)

5133.70

1299

1.208

(M2a–M1)0.07ns

3.049

.048–.051

.725

.958

.957

M2b:WB(T1)!

RS(T2)

5120.52

1299

1.207

(M2b–M1)8.09*

3.049

.048–.051

.757

.958

.957

w2¼Satorra-Bentler

scaledchi-square;SCF¼scalingcorrectionfactor;

Dw2¼scaledcorrectedchi-square

difference;Ddf¼difference

indegrees

of

freedom;*p5

.05;ns¼nonsignificant;RS¼role

stressors;WB¼workplace

bullying;T1¼Tim

e1;T2¼Tim

e2.

620 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

compared to the stability model. The different nested structural models arecompared using a scaled chi-square difference test, and critical values of thechi-square distribution relative to the degrees of freedom are taken asevidence of whether or not estimation of additional parameters is preferred(Joreskog, 1993).

The stability model showed acceptable fit to data, w2¼ 5134, df¼ 1302,RMSEA¼ .049, CFI¼ .958, NNFI¼ .957. However, the following model(M2a) with structural paths from the role stressors to exposure to workplacebullying did not show an improvement in model fit as compared to thestability model, w2scaled diff ð3Þ ¼ 0:07ns. Moreover, none of the structuralpaths from the role stressors at Time 1 showed any significant relationshipswith exposure to bullying at Time 2 (Table 3). However, the reverse model(M2b) with structural paths from exposure to workplace bullying at Time 1to the role stressors at Time 2 showed a significant improvement in model fitrelative to the stability model, w2scaled diff ð3Þ ¼ 8:09, p5 .05. The path fromworkplace bullying to role ambiguity showed a coefficient of .12, the path torole conflict a coefficient of .15, and the path to role overload a coefficient of.07, all three being significant (Figure 1). Thus, the results of the longitudinalstructural regression analyses did not support postulated relationshipsaccording to an individual-level model of role stressors being antecedents ofexposure to workplace bullying, whereas exposure to workplace bullying

Figure 1. Longitudinal structural model. WB1¼workplace bullying (T1); RA1¼ role

ambiguity (T1); RC1¼ role conflict (T1); RO1¼ role overload (T1); WB2¼workplace bullying

(T2); RA2¼ role ambiguity (T2); RC2¼ role conflict (T2); RO2¼ role overload (T2);

T1¼Time 1; T2¼Time 2; **p5 .01 (t+ 2.58), *p5 .05 (t+ 1.96).

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 621

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

was found to be a potent source of subsequent role stress for affectedindividuals.

Moreover, although demographical variables such as gender and agepreviously have been shown to only weakly relate to exposure to workplacebullying (cf. Aquino & Thau, 2009; Bowling & Beehr, 2006), the longitudinalsample employed in this study deviated significantly from the representativesample from which it was drawn in terms of both gender and age. Toinvestigate if such factors may have confounded the findings and led tospurious relations between the variables of interest (cf. Dormann, 2001), thefinal structural model was reestimated controlling for the possible influenceof gender and age, w2¼ 5410, df¼ 1387, RMSEA¼ .049, CFI¼ .956,NNFI¼ .954. However, the results showed no changes in the structuralrelationships between exposure to workplace bullying and the role stressors,thus indicating that the findings of the present study are not likely to be theresult of influence from this kind of variables that were not initially taken intoaccount.

DISCUSSION

The findings of the present study did not support the hypothesis ofindividual role stress being an antecedent condition of exposure toworkplace bullying. The structural regression analyses failed to identifyany significant individual-level relationships between the three role stressorsand subsequent exposure to bullying. Although theoretical frameworks andprior empirical research have assumed such causal relationships to exist onan individual level of analysis, no evidence of their validity in explainingexposure to workplace bullying were identified in the present sample. On theother hand, the hypothesis of individual role stress being a consequence ofexposure to workplace bullying received considerable support. Priorexposure to workplace bullying was found to significantly account forsubsequent variation in role ambiguity, role conflict, and role overload alike.Moreover, the strength of longitudinal relationships identified in the presentstudy also seems to be in accordance with patterns found in previous cross-sectional studies, with exposure to workplace bullying showing the strongestrelationship with role conflict, followed by role ambiguity and role overload.Hence, the assumption that individuals experiencing high levels of role stressmay elicit responses that can encourage victimization does not seem to hold.Thus, interpretations of cross-sectional findings along these lines ofarguments may be misleading and can lead to erroneous conclusions.

Relationships among individuals at work are influenced by severalfactors, such as liking, respect, the power balance between individuals, aswell as by role senders’ perceptions of how the focal person responds to rolepressure (Kahn et al., 1964). Role pressure may be exerted on individuals

622 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

not only as part of everyday interpersonal interaction within organizations,but also in the form of hostile and degrading interpersonal interaction.Legitimate role pressure may be employed to make the individual complywith organizational demands, and, if this pressure leads to the desiredchange in the focal person’s behaviour, there is no need to employ strongerpressure or to induce sanctions on behalf of the role sender. However, if rolesenders perceive that the individual ‘‘. . . does not seem to ‘hear’ what he isbeing told, perhaps they have to ‘shout’ all the louder . . .’’ (Kahn et al.,1964, p. 70). Thus, pressure in the form of bullying behaviour may beemployed to make the individual comply with expectations and demandsfrom others at work, leading to increased role stress in the focal person (cf.Ferris et al., 2007; Katz & Kahn, 1978).

Although theoretical frameworks have been proposed to account forindividual-level relationships between work-related factors and subsequentexposure to workplace bullying (cf. Bowling & Beehr, 2006), suchexplanations did not receive support in the present study. Nevertheless, acommon explanation found in the workplace bullying literature to accountfor such causal relationships is social interactionist theory (Neuman &Baron, 2003). According to this perspective, individually experienced stress(e.g., role stress) may be related to subsequent exposure to bullying throughits effect on the individual’s behaviour, a line of reasoning also found in thetarget-oriented model proposed by Bowling and Beehr (2006). If highlystressed individuals perform badly, violate established and accepted socialnorms of politeness, or annoy others in their work environment, theirbehaviour may evoke aggressive behaviour in both co-workers and super-visors. Retaliatory acts may then be employed to force the distressedindividual to comply, possibly resulting in the individual being exposed tobullying (Tedeschi & Felson, 1994). However, such a causal relationship isfar from perfect and actual relationships according to social interactionisttheory and target-oriented perspectives may not be readily observable. Asthe number of targets of bullying in randomly selected samples isnormally small (cf. Nielsen et al., 2009), and assuming that a largeproportion of these targets may be singled out on a highly arbitrary basis(cf. Leymann, 1992), the number of targets who have experienced priorhigh levels of role stress and subsequent exposure to bullying are likely tobe even fewer. Bullying will only occur within organizations that allowsuch behaviour to take place (Brodsky, 1976), and supervisors or fellowemployees may equally likely employ supportive strategies to counteractthe experienced stress as to engage in bullying behaviour towards theindividual. Thus, although it is possible that such a relationship doesexist for a small group of individuals, being able to detect causalrelationships in line with social interactionist theory and target-orientedapproaches in randomly selected samples seems quite unlikely.

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 623

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

What seems evident is that it is not necessarily the same individual whohas experienced prior high levels of ambiguous or conflicting workingconditions who will experience subsequent exposure to workplace bullying.It may very well be that it is other individuals in the work environment whoare experiencing high levels of work stress and who possibly projects theirtension and frustration onto a suitable scapegoat in the workgroup, therebybecoming perpetrators of bullying (cf. Baillien et al., 2009; Thylefors, 1987).Thus, it seems more likely that individually experienced role stress may berelated to perpetration of bullying than to exposure to bullying (Haugeet al., 2009). Such an assumption is also in line with theoretical frameworkslinking work-related factors and stressful working conditions to theexperience of negative emotions and to engaging in aggressive behaviourtowards others in the work environment (e.g., Spector & Fox, 2005).

Although, intuitively, the present findings may seem to contradictrelationships as proposed in theoretical frameworks on causality betweenwork-related factors and workplace bullying, the distinction betweenindividual-level and workgroup-level approaches also needs to be takeninto consideration. Most theoretical frameworks deduced from the workenvironment hypothesis seem to focus mainly on overall workgroup-levelprocesses, postulating causal relationships between stressful workingenvironments and bullying at the overall workgroup level, and assuming ahigher probability of individuals being exposed to workplace bullying instressful and poorly organized working environments as compared to lessstressful ones (cf. Leymann, 1996). So far, however, most studies haveanalysed and consequently interpreted such relationships at an individuallevel of analysis, with the possibility of thereby having committed anatomistic fallacy (cf. Hox, 2002). Individual-level interpretations of work-related factors as antecedents of exposure to workplace bullying may thus bemisleading. However, such individual-level relationships may rather relateto enactment of bullying (cf. Spector & Fox, 2005).

At an individual level of analysis of relationships between role stressorsand exposure to workplace bullying, ambiguous conditions such as lackingthe information necessary to perform one’s work satisfactorily may equallylikely be a consequence of being exposed to bullying as a cause of it.Likewise, being exposed to contradictory, incompatible, or excessivedemands from role senders may also be deliberate bullying strategiesemployed to put the focal person in a difficult position, resulting insubsequent perceptions of role conflict and role overload. However, thisdoes not mean that ambiguous or conflicting working conditions cannot actas antecedent conditions of bullying occurring in workgroups, merely thatexplanations for such causal relationships must take into account morecomplex and not readily observable relationships between the variables inquestion. The frustration experienced as a result of stressful working

624 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

conditions can give rise to aggressive behaviour in some individuals whomay then become perpetrators of bullying (cf. Spector & Fox, 2005).However, it may be quite arbitrary who is actually singled out as a target ofbullying within such stressful work environments on behalf of perpetrators,not to be explained by seeking its causes in the environmental experiences ofthe targets being exposed to bullying (cf. Leymann, 1992). Hence, it seemslikely that models based on analysis of individual-level data are toosimplistic to adequately map such complex and indirect processes as to whyindividuals are being exposed to workplace bullying, whereas modelsinvestigating relationships between work-related factors and bullying at anaggregated workgroup level may make it possible to detect relationships thatare not directly observable at an individual level (cf. Bliese & Jex, 2002).Such an assumption also seems to be more in accordance with the originalreasoning in the work environment hypothesis put forward by Leymann(1996) that within poor working environments characterized by a high levelof stress, the incidence of both targets and perpetrators of bullying is likelyto be higher than in more favourable environments. Nevertheless, gaining athorough understanding of what causes bullying in workplaces is certainlyimportant in terms of attempting to prevent and manage the problem.Future research is likely to benefit from adopting actor-oriented andmultilevel perspectives in this regard.

Methodological issues

The present study analysed relationships between role stressors andexposure to workplace bullying in a large sample of employees drawnfrom a representative sample of the Norwegian working population, thusincreasing the robustness and generalizability of the findings (see, e.g.,Nielsen & Einarsen, 2008, for a discussion of the use of conveniencesamples). By applying a structural equation modelling approach, thepresent study also addressed several issues raised in the literature relatingto longitudinal studies (cf. Brown, 2006; MacCallum & Austin, 2000).Although causal inferences cannot be proven within a longitudinal design,their plausibility can be established by ruling out alternative explanations(Zapf, Dormann, & Frese, 1996). Still, some caution needs to beexercised. Autoregressive models have been criticized for being too simplein terms of reflecting true change in constructs over time (cf. Chan, 2002),with researchers advocating the use of methods such as latent growthcurve modelling to overcome such shortcomings. However, as long-itudinal studies on workplace bullying are scarce, and as studiesinvestigating three or more waves are, to our knowledge, nonexistent inthis field of research, autoregressive modelling for two-wave data seemsjustifiable.

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 625

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

Furthermore, time lags in longitudinal studies and arbitrary measure-ment points will always be open to criticism. Too short a time lag maylead to the conclusion that no causal effect exists, whereas too long atime lag may lead to an underestimation of the true causal effect (Zapfet al., 1996). As the workplace bullying literature gives little indication ofwhen environmental factors may result in bullying, the validity of thepresent findings needs to be tested in future research. Nevertheless,because different time lags may have different effects and no one time lagis the only correct or true one, other time lags than the 2-year time lagapplied in this study should be investigated in order to more fullyunderstand the nature and effects of various time lags in relation tobullying (cf. Gollob & Reichardt, 1987). In addition, the substantialstability in exposure to workplace bullying over time may indicate thatthere is little variation left to predict in bullying at Time 2, and that thehigher stability in bullying, as compared to the role stressors, thereforemay account for the findings. However, as we have argued, thenonsignificant relationships between the role stressors and subsequentexposure to workplace bullying may equally likely be attributed to thelevel of analysis applied and to the distinction between exposure to andenactment of bullying. Thus, future research is likely to benefit fromtaking into account such perspectives when investigating environmentalcauses of workplace bullying.

CONCLUSION AND IMPLICATIONS

Given previous theorizing and empirical studies into the causes ofworkplace bullying, the findings and conclusions to be drawn from thepresent study may seem surprising at first glance. Previous research hasproposed work-related factors such as role stressors to be importantantecedents of exposure to workplace bullying, but the present findingsindicate that this does not hold true at an individual level of analysis.However, the finding that individuals exposed to bullying experiencesubsequent elevated levels of role stress should not be underestimated.Increased stress levels among individuals may, in turn, fuel a negativespiral in terms of the overall quality of the work environment, and thusact as an antecedent condition of bullying occurring in workgroups.Thus, although none of the role stressors investigated in the present studywere able to predict subsequent exposure to workplace bullying, thisfinding should not be interpreted to mean that role stressors or otherwork-related factors are unimportant in the development of bullying atwork, but merely that such relationships are not likely to be readilyobservable at an individual level of analysis. Investigations of work-related factors as causes of bullying at an aggregated workgroup level are

626 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

likely to more fully capture such complex processes leading up toindividuals being exposed to bullying and to individuals engaging inbullying of others, and thus to give a more nuanced picture of how suchprocesses unfold within different organizational contexts. In consequence,interventions are likely to be most effective if they focus on improving thecollective work environment. Not only will such an approach avoidvictim-blaming of exposed individuals, but organizations that adopt suchapproaches are also likely to benefit from a reduction in both overallstress levels among their employees and reduced levels of bullying.Although improving the overall work environment and countering thedevelopment of bullying should be the primary focus, leaders mustactively deal with each individual case of bullying that may occur. Theseverity of the situation as experienced by targets of bullying mustnevertheless be dealt with in an appropriate and satisfactorily manner,regardless of its causes.

REFERENCES

Agervold, M. (2009). The significance of organizational factors for the incidence of bullying.

Scandinavian Journal of Psychology, 50(3), 267–276.

Agervold, M., & Mikkelsen, E. G. (2004). Relationships between bullying, psychosocial work

environment and individual stress reactions. Work and Stress, 18(4), 336–351.

Anderson, J. C., & Gerbing, D. W. (1988). Structural equation modeling in practice: A review

and recommended two-step approach. Psychological Bulletin, 103(3), 411–423.

Aquino, K., & Thau, S. (2009). Workplace victimization: Aggression from the target’s

perspective. Annual Review of Psychology, 60, 717–741.

Baillien, E., Neyens, I., De Witte, H., & De Cuyper, N. (2009). A qualitative study on the

development of workplace bullying: Towards a three way model. Journal of Community and

Applied Social Psychology, 19(1), 1–16.

Beehr, T. A. (1995). Psychological stress in the workplace. London: Routledge.

Beehr, T. A., Walsh, J. T., & Taber, T. D. (1976). Relationships of stress to individually and

organizationally valued states: Higher order needs as a moderator. Journal of Applied

Psychology, 61(1), 41–47.

Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin,

107(2), 238–246.

Berkowitz, L. (1989). Frustration-aggression hypothesis: Examination and reformulation.

Psychological Bulletin, 106(1), 59–73.

Bliese, P. D., & Jex, S. M. (2002). Incorporating a multilevel perspective into occupational stress

research: Theoretical, methodological, and practical implications. Journal of Occupational

Health Psychology, 7(3), 265–276.

Bowling, N. A., & Beehr, T. A. (2006). Workplace harassment from the victim’s perspective: A

theoretical model and meta-analysis. Journal of Applied Psychology, 91(5), 998–1012.

Brodsky, C. M. (1976). The harassed worker. Toronto, Canada: Lexington Books.

Brown, T. A. (2006). Confirmatory factor analysis for applied research. New York: Guilford

Press.

Chan, D. (2002). Longitudinal modeling. In S. G. Rogelberg (Ed.), Handbook of research

methods in industrial and organizational psychology (pp. 412–430). Malden, MA: Blackwell.

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 627

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

Cohen, J. (1992). A power primer. Psychological Bulletin, 112(1), 155–159.

De Cuyper, N., Baillien, E., & DeWitte, H. (2009). Job insecurity, perceived employability and

targets’ and perpetrators’ experiences of workplace bullying. Work and Stress, 23(3),

206–224.

Dormann, C. (2001). Modeling unmeasured third variables in longitudinal studies. Structural

Equation Modeling, 8(4), 575–598.

Einarsen, S. (1999). The nature and causes of bullying at work. International Journal of

Manpower, 20(1/2), 16–27.

Einarsen, S. (2000). Harassment and bullying at work: A review of the Scandinavian approach.

Aggression and Violent Behavior, 5(4), 379–401.

Einarsen, S., Hoel, H., & Notelaers, G. (2009). Measuring exposure to bullying and harassment

at work: Validity, factor structure and psychometric properties of the Negative Acts

Questionnaire–Revised. Work and Stress, 23(1), 24–44.

Einarsen, S., Hoel, H., Zapf, D., & Cooper, C. L. (2003). The concept of bullying at work: The

European tradition. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and

emotional abuse in the workplace: International perspectives in research and practice

(pp. 3–30). London: Taylor & Francis.

Einarsen, S., Raknes, B. I., & Matthiesen, S. B. (1994). Bullying and harassment at work and

their relationship to work environment quality: An exploratory study. The European Work

and Organizational Psychologist, 4(4), 381–401.

Ferris, G. R., Zinko, R., Brouer, R. L., Buckley, M. R., & Harvey, M. G. (2007). Strategic

bullying as a supplementary, balanced perspective on destructive leadership. Leadership

Quarterly, 18(3), 195–206.

Gollob, H. F., & Reichardt, C. S. (1987). Taking account of time lags in causal models. Child

Development, 58(1), 80–92.

Graham, J. W. (2009). Missing data analysis: Making it work in the real world. Annual Review

of Psychology, 60, 549–576.

Hauge, L. J., Skogstad, A., & Einarsen, S. (2007). Relationships between stressful work

environments and bullying: Results of a large representative study. Work and Stress, 21(3),

220–242.

Hauge, L. J., Skogstad, A., & Einarsen, S. (2009). Individual and situational predictors of

workplace bullying: Why do perpetrators engage in the bullying of others? Work and Stress,

23(4), 349–358.

Hertzog, C., & Nesselroade, J. R. (1987). Beyond autoregressive models: Some implications of

the trait-state distinction for the structural modeling of developmental change. Child

Development, 58(1), 93–109.

Hoel, H., Zapf, D., & Cooper, C. L. (2002). Workplace bullying and stress. In P. L. Perrewe &

D. C. Ganster (Eds.), Historical and current perspectives on stress and health (Vol. 2,

pp. 293–333). Amsterdam: JAI Press.

Hox, J. (2002). Multilevel analysis: Techniques and applications. Mahwah, NJ: Lawrence

Erlbaum Associates, Inc.

Høstmark, M., & Lagerstrøm, B. O. (2006). Undersøkelse om arbeidsmiljø: Destruktiv atferd i

arbeidslivet. Dokumentasjonsrapport [A study of work environments: Destructive behaviours

in working life: Documentation report] (No. 44). Norway: Statistics Norway (SSB).

Jennifer, D., Cowie, H., & Ananiadou, K. (2003). Perceptions and experience of

workplace bullying in five different working populations. Aggressive Behavior, 29(6),

489–496.

Joreskog, K. G. (1993). Testing structural equation models. In K. A. Bollen & J. S. Long (Eds.),

Testing structural equation models (pp. 294–316). Newbury Park, CA: Sage.

Joreskog, K. G., & Sorbom, D. (1996). LISREL 8: User’s reference guide (2nd ed.). Chicago, IL:

Scientific Software International.

628 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

Joreskog, K. G., & Sorbom, D. (1999). PRELIS 2: User’s reference guide (3rd ed.). Chicago, IL:

Scientific Software International.

Kahn, R. L., Wolfe, D. M., Quinn, R. P., Snoek, D. J., & Rosenthal, R. A. (1964).

Organizational stress: Studies in role conflict and ambiguity. New York: Wiley.

Katz, D., & Kahn, R. L. (1978). The social psychology of organizations. New York: Wiley.

Keashly, L., & Harvey, S. (2005). Emotional abuse in the workplace. In S. Fox &

P. E. Spector (Eds.), Counterproductive behavior: Investigations of actors and targets

(pp. 201–235). Washington, DC: American Psychological Association.

Leymann, H. (1992). Fran mobbning til utslagning i arbetslivet [From mobbing to expulsion in

work life]. Stockholm: Publica.

Leymann, H. (1996). The content and development of mobbing at work. European Journal of

Work and Organizational Psychology, 5(2), 165–184.

MacCallum, R. C., & Austin, J. T. (2000). Applications of structural equation modeling in

psychological research. Annual Review of Psychology, 51, 201–226.

Matthiesen, S. B., & Einarsen, S. (2007). Perpetrators and targets of bullying at work: Role

stress and individual differences. Violence and Victims, 22(6), 735–753.

Neuman, J. H., & Baron, R. A. (2003). Social antecedents of bullying: A social interactionist

perspective. In S. Einarsen, H. Hoel, D. Zapf, & C. L. Cooper (Eds.), Bullying and emotional

abuse in the workplace: International perspectives in research and practice (pp. 185–202).

London: Taylor & Francis.

Nielsen, M. B., & Einarsen, S. (2008). Sampling in research on interpersonal aggression.

Aggressive Behavior, 34(3), 265–272.

Nielsen, M. B., Skogstad, A., Matthiesen, S. B., Glasø, L., Aasland, M. S., Notelaers, G., et al.

(2009). Prevalence of workplace bullying in Norway: Comparisons across time and

estimation methods. European Journal of Work and Organizational Psychology, 18(1),

81–101.

Notelaers, G., De Witte, H., & Einarsen, S. (2010). A job characteristics approach to

explain workplace bullying. European Journal of Work and Organizational Psychology, 19,

487–504.

Notelaers, G., & Einarsen, S. (2008). The construction and validity of the Short-Negative Acts

Questionnaire. Paper presented at the 6th international conference on Workplace Bullying,

Montreal.

Ortqvist, D., & Wincent, J. (2006). Prominent consequences of role stress: A meta-analytic

review. International Journal of Stress Management, 13(4), 399–422.

Rizzo, J. R., House, R. J., & Lirtzman, S. I. (1970). Role conflict and ambiguity in complex

organizations. Administrative Science Quarterly, 15(2), 150–163.

Spector, P. E., & Fox, S. (2005). The stressor-emotion model of counterproductive work

behavior. In S. Fox & P. E. Spector (Eds.), Counterproductive behavior: Investigations

of actors and targets (pp. 151–174). Washington, DC: American Psychological

Association.

Steiger, J. H. (1990). Structural model evaluation and modification: An interval estimation

approach. Multivariate Behavioral Research, 25(2), 173–180.

Tedeschi, J. T., & Felson, R. B. (1994). Violence, aggression and coercive actions. Washington,

DC: American Psychological Association.

Thylefors, I. (1987). Syndabockar. Om utstotning och mobbning i arbetslivet [Scapegoats: About

expulsion and bullying in working life]. Stockholm: Natur och Kultur.

Vartia, M. (2001). Consequences of workplace bullying with respect to the well-being of its

targets and the observers of bullying. Scandinavian Journal of Work, Environment and

Health, 27(1), 63–69.

Zapf, D. (1999). Organisational, work group related and personal causes of mobbing/bullying

at work. International Journal of Manpower, 20(1/2), 70–85.

ROLE STRESSORS AND EXPOSURE TO WORKPLACE BULLYING 629

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4

Zapf, D., Dormann, C., & Frese, M. (1996). Longitudinal studies in organizational stress

research: A review of the literature with reference to methodological issues. Journal of

Occupational Health Psychology, 1(2), 145–169.

Zapf, D., & Gross, C. (2001). Conflict escalation and coping with workplace bullying: A

replication and extension. European Journal of Work and Organizational Psychology, 10(4),

497–522.

Original manuscript received December 2009

Revised manuscript received March 2010

First published online September 2010

630 HAUGE, SKOGSTAD, EINARSEN

Dow

nloa

ded

by [

Uni

vers

ity o

f C

hica

go L

ibra

ry]

at 0

4:04

19

Oct

ober

201

4


Top Related