Workplace bullying: A risk control perspective
MetadataShow full item record
Workplace bullying is an omnipresent phenomenon in contemporary workplaces (Nielsen, Matthiesen, & Einarsen, in press). With its negative consequences for victims, bystanders and the socio-economic fabric of organisations, it is an important psychological, sociological and economical hazard that needs to be firmly addressed. Several countries, such as Norway, Sweden, Finland, Belgium and France, have adopted laws explicitly banning it from workplaces. Some European countries have integrated this hazard into their general working environment acts as a result of applying European legislation that underlines the need for a risk assessment approach in the workplace (EEC, 1989). In this thesis, a risk control cycle (Cox, 1993) adapted from safety and health sciences (Karanika-Murray, Antoniou, Michaelides, & Cox, 2009) is introduced as an obvious strategy to systematically manage the problem of bullying at work. Up until now, empirical data concerning the effectiveness of interventions in connection with bullying are a very early stage (Salin, 2009) or even inconclusive (Leka et al., 2008). Some of the key deficiencies are of a methodological nature. Furthermore, bullying is a somewhat intangible phenomenon, hence difficult to recognise and capture (Leka, Vartia, et al., 2008). This thesis aims to contribute to the feasibility of the risk control cycle by focusing on the very first step of risk control, which is the identification of the hazard and its antecedents in order to obtain a trustworthy benchmark or reference point. This is a condition sine qua non for the next steps in the control cycle, i.e. the assessment of risk, the design of control measures and the evaluation of intervention strategies to counteract workplace bullying. In particular, identifying the targets of severe bullying, assessing risk groups and mapping job characteristics associated with severe bullying are central to achieving this aim, and they are therefore the focus of this thesis, which consists of four studies. The aim of the first study was to empirically explore the nature of workplace bullying. In particular, the objective of the study was to estimate the number of target groups of workplace bullying. The responses to the Negative Acts Questionnaire (Einarsen & Raknes, 1997) were therefore analysed in a large heterogeneous sample. A latent class cluster analysis (Magidson & Vermunt, 2004; Vermunt & Magidson, 2002) distinguished six different exposure groups that differed with respect to the nature and the frequency of the reported negative acts. In particular, it was found that approximately 3% of the sample was a target of severe bullying. The construct and criterion validity of these six groups emphasises the distinctiveness of the groups by showing significantly different scores for well-being and strain. Hence, bullying is not an either-or phenomenon, rather a hazard that comes in various shapes and forms. When designing interventions, the different clusters yield valuable clues as to what kind of specific approach an organisation can take. After having established that latent class modelling is a valuable way of detecting target groups of bullying, the second study aimed to assess the risk groups in relation to workplace bullying. For this purpose, a multinomial regression model was applied to a large and heterogeneous sample of Flemish-speaking employees. Especially public servants, bluecollar workers and employees in the food industry displayed substantially higher risks of being targets of severe bullying. Employees between the ages of 35 and 54 and employees who work in manufacturing industry had also a higher risk of being a target of severe bullying. The likelihood of being a target of severe bullying did not differ between those on temporary and permanent contracts, nor between male and female workers. Nonetheless the necessity to identify the nature and prevalence of the phenomenon and its risk groups, the design of interventions may also benefit from a better understanding of the antecedents of workplace bullying. This may be accomplished by investigating the psychosocial working environment or the design of work. The present thesis will therefore also investigate job characteristics that may be likely antecedents of bullying. In exploring how job characteristics may act as antecedent factors of exposure to severe bullying, this thesis aimed to investigate whether job stress, and thus the experience of stress stemming from the design of jobs, may lead to reports of workplace bullying (Baillien, Neyens, De Witte, & De Cuyper, 2009). In study three, it was investigated whether job strain as conceived by the Job Demand Control Model (Karasek, 1979) can account for the likelihood of being severely bullied in a large and heterogeneous sample. In addition, endeavours were made to establish the level of experienced demand and control at which this likelihood increased significantly (Hoel, Zapf, & Cooper, 2002). The results underlined that experiencing strain is related to be being bullied. Moreover, the risk of being classified as a target of severe bullying is particularly high when demands are very high and control is low, and especially so when demands are very high and control is almost absent. In order to complement the strain hypothesis, the last study aimed to identify job characteristics that may be associated with exposure to bullying. The Vitamin modelӳ exhaustive list of environmental features was used to formulate the various research hypotheses (Warr, 1987, 1990). Participatory decision-making, role conflicts and environmental clarity acted as antecedent factors for exposure to bullying, based upon a binomial regression model in which several social demographical variables were controlled for. From a prevention perspective, when trying to control exposure to workplace bullying, it is of the utmost importance to design jobs in such a way that tasks and responsibilities are clear and not mutually contradictory, while offering opportunities for control. All in all, these four studies contribute to the establishment of a trustworthy and necessary baseline for accomplishing the remaining steps in the risk control cycle. Different target groups have been identified, risk groups have been assessed and some important job characteristics have been documented that may be antecedent factors in relation to workplace bullying. The first paper also showed significant and substantial relationships between exposure to bullying and measures of health and well-being. This is important for the next steps of the control cycle because it demonstrates the urgency of managing the hazard. Subsequently, interventions can be designed, evaluated and monitored to achieve the goal of risk control, i.e. reducing, and if possible, banning bullying from our workplaces.
Paper I: Work and Stress 20(4), Notelaers, G.; Einarsen, S.; De Witte, H. & Vermunt, J. K., Measuring exposure to bullying at work: The validity and advantages of the latent class cluster approach, p. 288-301. Copyright 2006 Taylor & Francis. Full text not available in BORA. The published version is available at: http://dx.doi.org/10.1080/02678370601071594.Paper II: Industrial Health 49, Notelaers, G.; Baillien, E.; Vermunt, J. K.; De Witte, H. & Einarsen, S., Exploring risk groups and risk factors for Workplace Bullying. Copyright 2010 National Institute of Occupational Safety and Health. Full text not available in BORA. Accepted version. The published version is available at: http://dx.doi.org/10.2486/indhealth.MS1155Paper III: Economic and Industrial Democracy, 2011, Notelaers, G.; Baillien, E.; De Witte, H.; Einarsen, S. & Vermunt, J. K., Testing the strain hypothesis of the demand control model to explain severe bullying at work. Submitted version. Full text not available in BORA.Paper IV: European Journal of Work and Organizational Psychology 18(4), Notelaers, G.; De Witte, H. & Einarsen, S. , Workplace bullying: a job characteristics approach, p. 487-504. Copyright 2010 Taylor & Francis. Full text not available in BORA. The published version is available at: http://dx.doi.org/10.1080/13594320903007620.