The moderating role of perceived organisational support in the relationship between workplace bullying and turnover intention across sectors in South Africa

Note: This article was republished with the correct affiliation for Lena-Mari van Schalkwyk Orientation: Because workplace bullying has detrimental consequences on the profitability, work quality and turnover intention of organisations, this phenomenon should be addressed. Perceived Organisational Support (POS) was explored since factors such as role clarity, job information, participation in decision-making, colleague support and supervisory relationships might act as buffers against workplace bullying, subsequently influencing the turnover intention of the organisation.


Introduction Key focus of the study
Despite the increasing demands of the global world, which constantly generates new technological aspirations, work relationships are still recognised as a central component of workplaces worldwide.Workplace bullying by either an individual (Sperry, 2009) or multiple perpetrators (Namie & Namie, 2009) is increasing in the workplace.Experiences of bullying behaviour in the workplace are four times more prevalent than illegal forms of workplace harassment such as sexual harassment (Namie, 2007).Furthermore, bullying in the workplace is recognised internationally (Baillien, Neyens, De Witte & De Cuyper, 2009;Einarsen, Hoel & Notelaers, 2009;Lutgen-Sandvik & Sypher, 2009;Ortega, Høgh, Pejtersen & Olsen, 2009) and nationally (Pieterson, 2007;Wright, 2008) as a relevant and destructive phenomenon.
Most international studies have focused on the development and measurement of workplace bullying (Baillien et al., 2009;Einarsen et al., 2009;Namie & Namie, 2009), characteristics of workplace bullying (Lutgen-Sandvik & Sypher, 2009;Ortega et al., 2009 ) and the different sectors experiencing workplace bullying (Bilgel, Aytac & Bayram, 2006;Bloisi & Hoel, 2008;Mathisen, Einarsen & Mykletun, 2008).This study will focus on two forms of workplace bullying, viz., experiences of bullying behaviour from superiors and experiences of bullying behaviour from colleagues.Furthermore, the study investigated whether perceived organisational support acts as a buffer against workplace bullying and endeavoured to show the relationship between workplace bullying and employee turnover intention.

Background to the study
In the context of Industrial Psychology, international research has shown a growing interest in workplace bullying (Agervold, 2007).Conversely, workplace bullying in South Africa is still in its infancy.This aggressive behaviour affects personal and professional relationships throughout an individual's lifespan (Lewis, Coursol & Wahl, 2002) and if not recognised by organisations as a workplace phenomenon, it will increase.Workplace bullying occurs between superiors and their staff, and more recently, horizontally, which entails bullying among colleagues (Lewis & Sheehan, 2003).However, research has shown that abusive superiors have a more profound effect on employee commitment than experiencing abusive behaviour from colleagues (Koonin & Green, 2004;McCormick, Casimir, Djurkovic & Yang, 2006).
Nevertheless, if an organisation acts in a professional manner by providing the necessary support to workplace bullying targets, these supportive measures may enable employee coping mechanisms to deal with abusive behaviour in a constructive manner (Quine, 1999).Similarly, many researchers report that a lack of support is central to the inability of workplace bullying targets to cope with this phenomenon (Lewis & Orford, 2005;Leymann & Gustafsson, 1996;Matthiesen, Aasen, Holst, Wie & Einarsen, 2003).Then again, workplace bullying is under-reported and becoming an increasingly silent epidemic because of a lack of perceived organisational support and a fear of retribution (Koonin & Green, 2004;Lewis et al., 2002;MacIntosh, 2005;Strandmark & Hallberg, 2007).Consequently, the responsibility to cope with and report workplace bullying experiences does not rest only with the employee.It is also the responsibility of the organisation to protect employees against such actions in the workplace.Namie (2000) acknowledges that 96% of co-workers are aware of the bullying target's situation, suggesting that bullying is not a workplace secret.Even if co-workers do not witness workplace bullying, 87% of targets tell co-workers about their experiences.However, instead of protesting the bullying behaviour, employees on the sidelines often rally in support of the perpetrator; this is usually done because of a fear of punishment and as a way of self-protection.Because of this bullying, targets are incapable to form supportive coalitions with colleagues (Namie, 2000).
Workplace bullying is linked to various physical and psychological costs for the target, as well as organisational and social costs for the workplace.Consequences relating to the organisation include compensation for medical expenses (Bassman, 1992), and reduced productivity (Einarsen, Hoel, Zapf & Cooper, 2003) due to poor work performance.According to Einarsen et al. (2003) and Rayner and Keashly (2005), low-quality work, reduced productivity, high staff turnover and increased absenteeism are among the indirect costs to organisations.Targets contemplating strategies to cope with this destructive phenomenon found resignation to be an appropriate solution (Björkqvist, Österman & Hjelt-Bäck, 1994;Davenport, Schwartz & Elliott, 1999;Namie, 2000;Quine, 1999).

Research objectives
The primary objectives of this study were to determine whether perceived organisational support (role clarity, job information, participation in decision-making, colleague support and supervisory relationships) moderate the relationship between workplace bullying (by superiors and colleagues) and turnover intention.

Workplace bullying
Workplace bullying can be regarded as an element of aggressive behaviour that manifests in interpersonal work relationships between two individuals or between an individual and a group (Zapf & Einarsen, 2001).Rothmann and Rothmann (2006, p. 14) define bullying as 'harassing, offending, and socially excluding someone at work to such a level that these actions negatively affect a person's work tasks'.Thus, the label of the bully can be applied when a particular activity, interaction, behaviour or process has occurred repeatedly, frequently and over a period of time (Wood, 2008).Bullying is an escalating process where the victim experiences systematic negative social acts that lead to inferiority (Einarsen et al., 2003).
According to Rayner (1997), different types of bullying can be categorised as follows: threat to professional status (belittling opinion, public professional humiliation, accusation regarding lack of effort); threat to personal reputation (name-calling, insults, intimidation, devaluing); isolation (preventing growth opportunities, physical or social isolation, withholding information); overwork (undue pressure, impossible deadlines, unnecessary disruptions); and destabilisation (failure to give recognition, assigning of meaningless tasks, removal of responsibility, repeated reminders of blunders and setting the victim up to fail).Namie and Namie (2003) suggested that 82% of employees who had been bullied left their workplaces, 38% for health reasons and 44% because they were victims of a low performance appraisal manipulated by a bullying superior to show them as incompetent.According to Watkins (2007), a person who experiences persistent intimidation might learn to expect bullying behaviour from others and develop a pattern of compliance with the unfair demands of those he or she perceives as stronger.Furthermore, Baillien et al. (2009) advised that future research should focus on the distinction between different perpetrators of bullying in the organisation, such as superiors and colleagues.
Hypothesis 1: Workplace bullying by superiors will be more prevalent in organisations than workplace bullying by colleagues.
Bullying behaviours also exist because of a 'white wall of silence', where the superior often defends the perpetrator (Murray, 2007).Consequently, the superior can be the bully or even be the second degree perpetrator in a bullying situation where he or she defends the bully.According to IOMA (July, 2008), witnesses to workplace bullying believed in 43% of cases that the perpetrator of bullying had the support of one or more senior managers when harassing a victim.Similarly, Longo and Sherman (2007) suggest that superiors manipulate behaviour and often protect the bully instead of the victims.As a result, reporting bullying behaviour is often an unsatisfactory solution, especially since superiors are more likely to be associated with being the 'support from the organisation'.
According to Leymann (1990), bullying exists in organisations characterised by deficiencies in work design, leadership and negative social climates.It is argued that where managers avoid taking charge or involving themselves with work and stress or interpersonal conflicts and tensions, there is a breeding ground for bullies (Leymann, 1996).Consequently, it can be expected that those exposed to bullying will experience their immediate superior as an abusive and tyrannical leader.
In approximately a third of incidents, victims identified their colleagues or peer groups as the perpetrators, although bullying by colleagues is also interlinked with bullying by superiors.Conflict might be the origin of aggressive behaviour between colleagues, and potentially escalates into bullying when the behaviour becomes deliberate and purposeful (Strandmark & Hallberg, 2007).New or younger employees in the organisation are considered to be particularly vulnerable, as are ethnic minorities, owing to a lack of knowledge regarding their rights and the regulations of the workplace (Westhuses, 2004).Intimidation and blame in the organisation creates mutually held fears about future job security among employees (Vaez, Ekberg & LaFlamme, 2004).Job insecurity creates a climate of rivalry when employees see their colleagues as potential rivals for jobs.This may cause feelings of competition and suspicion, factors that are known to be associated with workplace bullying (Bjorkqvist et al., 1994).
Manipulative and inappropriate behaviour from colleagues has led to one in five employees quitting their jobs, creating a substantial staff retention problem for employers.A further 23% of employed people have raised complaints of bullying in the workplace but in two out of three cases, issues remain unresolved or the complaints have failed to affect the bullying campaigns in any way (Mason, 2010).

Bullying and perceived organisational support
According to Rhoades and Eisenberger (2002), perceived organisational support (POS) is founded on the assumption that employees form opinions regarding the extent to which an organisation values their contributions and cares about their well-being; moreover, less workplace bullying has been reported in organisations with a supportive (accommodating, trustworthy and caring) climate (Baillien, Neyens & De Witte, 2004).
Targets as well as observers of workplace bullying suffer from an ill-conditioned work environment (Sheehan, Ramsay & Patrick, 2000) and these ill-conditioned work environments with role conflict, a lack of participation in decision-making processes and a lack of support from superiors emancipate a bullying culture (Quine, 2001).
Role conflict or role ambiguity manifests in work environments where bullying exists.Einarsen, Raknes and Matthiesen (1994) highlight the fact that employees who are uncertain about their job expectations will experience a lack of support from the organisation.Moreover, targets of bullying rated the work environment more negatively, particularly with respect to the experience of role ambiguity (Vartia, 1996;Zapf, 1999).Therefore, role clarity can buffer the impact of workplace bullying through providing a sense of perceived support from the organisation.Furthermore, one the one hand, clear evidence indicates that negative acts by superiors and colleagues, such as imposing demeaning tasks, excessive monitoring, excessive criticism, withholding of information and exclusion thereof, represent workplace bullying (Lewis & Gunn, 2007).On the other hand, factors shown to buffer the manifestation of workplace bullying can be justified by supportive management structures (Cummings & Worley, 1997), and worker involvement and participation in decisionmaking (Marchington, 1995).In addition, studies conducted in the UK and Scandinavia have shown that many employees from both public and private sector organisations are frequently subjected to physical and verbal abuse by their colleagues and/or superiors (Adams, 1992;Leymann, 1990;Randall, 1997;Wilson, 1991).A lack of social support from colleagues in the work environment places targets in a more vulnerable position and so they become an easier target for bullies (Sheehan et al., 2000).
The aforementioned information clearly demonstrates that without a supportive culture in the organisation, workplace bullying will lead, directly or indirectly, to targets in the organisation to have an increased intention to resign (Djurkovic et al., 2004;Keashly, 2001;Quine, 2001).
Hypothesis 2: A negative relationship exists between workplace bullying by superiors and perceived organisational support (role clarity, job information, participation in decision-making, colleague support and supervisory relationships).
Hypothesis 3: A negative relationship exists between workplace bullying by colleagues and perceived organisational support (role clarity, job information, participation in decision-making, colleague support and supervisory relationships).

Bullying and turnover intention
Workplace bullying has been found to be a significant predictor of turnover (Begley, 1998), which incurs substantial costs for the organisation (Waldman, Kelly, Arora & Smith, 2004).
Hypothesis 4: A positive relationship exists between workplace bullying by superiors and turnover intention.
Hypothesis 5: A positive relationship exists between workplace bullying by colleagues and turnover intention.
Workplace bullying has widespread negative effects on organisations because it affects not only the targets but also those who witness bullying behaviour.Because of employers' costs associated with bullying, such as: productivity loss, costs regarding interventions by third parties, turnover, increased sick-leave, workers' compensation and disability claims and legal liability -employers should logically be motivated to stop workplace bullying (Hoel & Einarsen, 2010).

Perceived organisational support and turnover intention
Researchers have focused significant attention on the concept of POS as a key predictor of turnover intention (Maertz, Griffeth, Campbell & Allen, 2007).Hui, Teo and Lee (2007), examined both turnover intention and POS and concluded that POS was negatively related to thoughts of leaving the organisation.Conversely, POS was also positively related to staying with the organisation.POS was negatively related to thoughts about quitting a job as a result of negative acts in the workplace.Similarly, Kinnunen, Feldt and Makikangas (2008) found that POS was negatively related to the likelihood of leaving an organisation and the frequency of thoughts about leaving the organisation.
Accordingly, 'a social network and social support are valuable resources that not only enable individuals to cope with a wide variety of extant stressors but may also facilitate proactive coping efforts' (Aspinwall & Taylor, 1997, p. 421).In a longitudinal study of manufacturing workers, Moore, Grunberg and Greenberg (2004) found that greater role clarity was significantly associated with less turnover intention.Moreover, role clarity creates a sense of purpose for employees, leading to the retention of employees by the organisation (Sümer & Van Den Ven, 2008).In addition, a variety of abusive supervisory behaviours (Zellars, Tepper & Duffy, 2002) and a lack of participation in decision-making intentions (Allen et al., 2003) have been identified as situations where the only solution lies in targets quitting their jobs.Thus, Eisenberger, Huntington, Hutchison and Sowa (1986) argue that targets who perceive greater support from their employing organisation would be more likely to feel obligated to 'repay' the organisation (Shore & Wayne, 1993).
To this end, the goal of this study is to add to the research on the existence of workplace bullying in the South African context and how targets of bullying perceive support from the organisation where such a culture prevails.The impact of this phenomenon on turnover intention in organisations should create encouragement to generate transformation.Hypothesis 6: A negative relationship exists between perceived organisational support (role clarity, job information, participation in decision-making, colleague support and supervisory relationships) and turnover intention.
Hypothesis 7: Perceived organisational support (role clarity, job information, participation in decision-making, colleague support and supervisory relationships) plays a moderating role in the relationship between workplace bullying (superiors and colleagues) and turnover intention.

The potential value-add of the study
Avoidance by employers and employees to acknowledge the problem of workplace bullying hinders the awareness of the detrimental effects of such bullying on the world of work.By understanding the impact of bullying, organisations might be stimulated to buffer workplace bullying with sufficient support, such as role clarity, sufficient job information, joint decision-making and support from both the supervisors and colleagues of the targets.
International research (Baillien et al., 2009;Einarsen et al., 2009;Lutgen-Sandvik & Sypher, 2009;Namie & Namie, 2009;Ortega et al., 2009) has established the significance of workplace bullying worldwide.Consequently, this phenomenon cannot be ignored in the South African context.South Africa is still in the exploration phase of how workplace bullying affects a diverse developing country (Pieterson, 2007;Wright, 2008).Hence, this study will not only contribute to a better understanding of workplace bullying trends in South Africa and its relationship with employee turnover intention, but will also provide for preventive measures to minimise the risk of bullying behaviour in South African workplaces.
Contemporary research is essential in determining the influence of workplace bullying on the intention to leave an organisation.However, greater emphasis on the effects of workplace bullying creates more awareness so that organisations and employees can counteract its occurrence.Thus, organisations can be alert and prepared to target workplace bullying through interventions, policies and, as this study will reveal, perceived organisational support, in order to counteract this issue.

What will follow
In the sections to follow the research approach, research participants, measuring instruments, research procedure and statistical analyses are discussed.The results of the study are presented, followed by a discussion of the results.

Research approach
This article follows a quantitative research approach with a cross-sectional field survey.This approach is conducted by means of questionnaires to measure workplace bullying, turnover intention and POS at a single point in time (Trochim & Donnelly, 2007).

Research participants
The sample group is represented by an availability sample of 13 911 participants gathered over a spectrum of nine provinces and five sectors.The first table presents the biographical characteristics of the participants.
The sample (N = 13 911) embodies a variety of sectors in the South African industry (academic, financial, government, manufacturing, mining and other), in which the mining industry is represented by 5197 (37.4% of the total sample) participants and the academic environment has the lowest representation, with 209 (1.5%) participants.Afrikaansspeaking (5486) and Xitsonga-speaking (5486) participants make up most of the sample group with 78.8%, while 81 isiNdebele participants (0.6%) make up the smallest proportion of the language category.There are almost twice as many men, with 9158 (65.8%) participants, as women.

Measuring instruments
A biographical questionnaire was administered in order to document the socio-demographic differences of the participants.Characteristics on this questionnaire included gender, age, race, language, marital status, educational level, geographical distribution and industry.
The South African Employee Health and Wellness Survey (SAEHWS) was administered to gather the data.The SAEHWS is a self-report instrument based on the dualprocess model of work-related well-being (Rothmann & Rothmann, 2006), developed by Demerouti, Bakker, Nachreiner and Schaufeli (2001), and is rooted in the assumption that employees' perceptions and experiences represent imperative information regarding the wellness climate in an organisation.The validity of the factor structures of the SAEHWS is equivalent for different ethnic groups and organisations and is therefore culturally sensitive with no bias against any cultural group (Rothmann & Rothmann, 2006).South African norms were also developed and Rothmann and Rothmann (2006) reported that the reliability of the SAEHWS was satisfactory, with a Chronbach's Alpha coefficient above 0.70.
For the purpose of this study the following sections from the SAEHWS were used: Workplace bullying was measured in terms of bullying by superiors and bullying by colleagues to determine the main culprits.The scale for workplace bullying by superiors consisted of a 1-4 ('never' to 'always') Likert scale range with 12 items asking questions such as: 'How often do you experience unpleasant personal remarks from your superiors?' (α = 0.87).The scale for workplace bullying by colleagues is also determined by a 1-4 Likert scale ('never' to 'always') where responses to 12 items can be evaluated, with questions such as: 'How often do you feel that your colleagues are spreading unfair rumours about you?' (α = 0.86).
To determine the role of POS in this study, it was measured, for the sake of accuracy, by utilising the sub-facets of the stated definition (role clarity, job information, participation in decision-making, colleague support and supervisory relationships).Therefore, questions will be representative of each sub-facet.All sub-facets will be measured on a Likert scale 1-4 ('never' to 'always') range with three items, including questions relating to the various sub-facets, such To determine the cost to the company, turnover intention was explored.The turnover intention was rated on a 1-6 Likert scale ('strongly disagree' to 'strongly agree') with five items, where questions like: 'I would be very happy to spend the rest of my career in this organisation' were asked (α = 0.86).

Research procedure
The data were gathered in collaboration with AfriForte over a period of three years from 2007 to 2010.The data were collected from all nine provinces and specified sectors.
Participants were requested to follow a link received via e-mail and to complete the computerised questionnaire online.A letter of informed consent was completed by all participants and confidentiality was maintained throughout the research process.

Statistical analysis
The statistical analysis was carried out by means of the SPSSprogramme (SPSS Inc., 2009).Descriptive statistics (e.g.means, standard deviations, skewness and kurtosis) were used to analyse the data.Cronbach Alpha coefficients were used to assess the internal consistency, homogeneity and unidimensionality of the measuring instruments (Clark & Watson, 1995).The Eigen values and screen plot were studied to determine the number of extracted factors.A principal components analysis with a direct Oblimin rotation was conducted in the case where factors were related (r > 0.30).
A principal component analysis with a Varimax rotation was used if obtained factors were not related (Tabachnick & Fidell, 2001).
To specify the relationship between the variables, Pearson product-moment correlation coefficients were used in the case of normal distribution, and Spearman product-moment correlation coefficients for skew distributions.In terms of statistical significance, it was decided to set the value at a 95% confidence interval level (p ≤ 0.05).Effect sizes (Steyn, 1999) were used to decide on the practical significance of the findings.The practical significance of correlation coefficients was set with a medium effect (p ≥ 0.30) and a large effect (p ≥ 0.50).
Logistic regression analysis was conducted to determine the proportion of variance in the dependent variable (turnover intention) that is predicted by the independent variable (workplace bullying).A hierarchical regression analysis, as recommended by Aiken and West (1991), was performed in order to determine the moderating effect of the moderator variable (POS) on the relationship between the independent variable (workplace bullying) and the dependent variable (turnover intention).To provide a clearly interpretable interaction term and to reduce multicollinearity, the variables were standardised.In Step 1, the independent variable (workplace bullying) was regressed with the dependent variable (turnover intention).In the following steps, the dimensions of the moderator variable (POS, i.e. role clarity, job information, participation in decision-making, colleague support and supervisory relationship) were entered.
The order of entering these variables depended on the strength of the correlations between the dimensions of the moderator variable (perceived organisational support) and the dependent variable (turnover intention).In the final step, the interaction term (workplace bullying × moderator) was added, and a moderating effect was confirmed if the interaction term was statistically significant and if explained variance (R 2 ) was significantly increased (p < 0.05).

Descriptive statistics, internal consistencies and correlations
Table 2 provides the descriptive statistics, internal consistencies (Cronbach alpha coefficients) and correlations between workplace bullying, POS (role clarity, job information, participation in decision-making, colleague support and supervisory relationships) and turnover intention.
Table 2 represents the satisfactory Cronbach alpha coefficients obtained for all the scales which were higher than the guideline of a > 0.70 (Nunnally & Bernstein, 1994).
All the items are normally distributed except for bullying by superiors and colleagues, role clarity and job information.
Therefore, for these constructs the Spearman productmoment correlations will apply.Pearson product-moment correlations were used for all the other scales.
Table 3 indicates that bullying by superiors shows a positive statistically and practically significant correlation (with a large effect) with bullying by colleagues, and a negative statistically and practically significant correlation (with a large effect) with supervisory relationships.Also, negative statistically and practically significant correlations were found (with a medium effect) between bullying by superiors and work role clarity, job information and participation in decision-making.A positive statistically significant correlation was found between bullying by superiors and turnover intention.
Bullying by colleagues is negatively correlated with all constructs with a statistically significant relationship; however, practically significant correlations with a medium effect were found with colleague support and supervisory relationships.A positive statistically significant correlation was found between bullying by colleagues and turnover intention.
Role clarity shows a positive statistical significance with all constructs and a practical significance (with a large effect) with job information, supervisory relationships and participation in decision-making.A practical significance (with a medium effect) was found between role clarity and colleague support, whereas a negative practically significant correlation was shown for turnover intention.Job information shows a positive statistically and practically significant correlation with all constructs: supervisory relationships and participation in decision-making (large effect) and colleague  support, and a negative statistically and practically significant correlation with turnover intention (medium effect).
To address the last objective of this research study, a hierarchical regression analysis was conducted in order to determine if POS acts as a moderator in the relationship between workplace bullying (either by superiors or by colleagues) and turnover intention.
In step one of the hierarchical regression, role clarity was regressed on turnover intention (β = -0.32,p = 0.00) producing a significant model.The entering of bullying by superiors in step two produced a significant model (β = 0.18, p = 0.00).In the third step the interaction term was entered, producing a significant model (β = 0.03, p = 0.00).Thus, role clarity, as one of the defined sub-facets of POS, proves to play a significant role in the relationship between workplace bullying by superiors and the tendency to leave the organisation.A target who has clarity on his or her role might consider staying in the organisation, despite workplace bullying which would usually be serious enough to lead to resignation.
Figure 1 shows that when bullying by superiors manifests, targets who perceive less support from their organisations (in experiencing role ambiguity) have a much higher tendency to leave the organisation.However, it can also be seen that when targets' experience of role clarity increases, those targets being bullied by their superiors will, in comparison with those employees who experience low levels of support, be more inclined to stay in the organisations despite the bullying culture.
In step one of the hierarchical regression, participation in decision-making was entered and a significant model was produced (β = -0.37,p = 0.00).Bullying by superiors was entered in step two which produced a significant model (β = 0.14, p = 0.00).In the third step the interaction term was entered, producing a significant model (β = 0.03, p = 0.00).This indicates that participation in decision-making, as one of the defined sub-facets of POS, plays a significant role in the relationship between workplace bullying by superiors and the tendency to leave the organisation.If the targets of bullying by superiors feel that they can participate in the decision-making processes of the organisation, this indicates that their perceived support from the organisation is intact.When these targets are bullied by their superiors, having perceived support in the form of participation means that they do not leave the organisation, despite being bullied.On the other hand, when experiencing bullying from a superior while also not being informed or included in the decisionmaking processes, the target's final counteraction towards the bullying culture will be to leave the organisation.
Figure 2 shows that when targets of bullying by superiors experience POS through participation in decision-making, their tendency to leave the organisation will be minimised.However, it can also be seen that when targets experience exclusion from decision-making processes together with being bullied by their superiors, they will be more inclined to leave the organisation because of a lack of support from the organisation.
Supervisory relationship was entered in step one of the hierarchical regression and a significant model was produced (β = -0.32,p = 0.00).Bullying by superiors was entered in step two which produced a significant model (β = 0.14, p = 0.00).
In the third step the interaction term was entered, producing a significant model (β = 0.06, p = 0.00).This implies that relationships with a superior play a significant role in the relationship between bullying by superiors and turnover intention.When the targets experience healthy relationships with their superiors they will feel supported by their superiors and, therefore, by the organisation.When bullying by superiors manifests but the perceived support through supervisory relationships exists, the targets of the bullying phenomenon will not be inclined to leave the organisation.However, when the employees have poor working relationships with their superiors while also experiencing bullying from them, the targets will be much more driven to resign than to fight an overwhelming battle.
Figure 3 indicates that when employees have insufficient support from their organisations due to unhealthy relationships with their superiors, workplace bullying by superiors can easily emerge.Conversely, when employees' POS exists with the focus on healthy relationships with their superiors, being bullied by their superiors will not be enough to drive the targets of bullying by superiors to find their solution in absenteeism leading to resignation.
The aforementioned hierarchical regression analysis reveals that when employees are bullied by superiors, role clarity, participation in decision-making and supervisory relationships (sub-facets of POS) will moderate the relationship between superior bullying and turnover intention.There seem to be no moderating relationships when being bullied by colleagues.

Discussion
The main purpose of this study was to explore bullying (by superiors and colleagues), perceived support by employees in the organisation and the influence of workplace bullying on the turnover intention.Therefore, POS (role clarity, job information, participation in decision-making, colleague support and supervisory relationships) was investigated as a possible moderator in the relationship between workplace bullying, either by superiors or colleagues, and the intention to leave the organisation.As mentioned in the literature review, little research has been done in the South African context concerning the perpetrators of bullying, the relationship between workplace bullying and turnover intention, and the buffering effect of POS on this relationship.
The internal consistency reported that all the constructs were reliable, with a coefficient above 0.7.The reliability coefficients varied between 0.82 and 0.87.All the items were normally distributed except for bullying by superiors and colleagues, role clarity and job information.Consequently, the Spearman product-moment correlations were utilised.
Pearson product-moment correlations were used for all the other scales as their distributions were normally distributed.These correlations reported on the type of the relationship and the effect sizes between the variables.
When interpreting the results, the first hypothesis, that bullying by superiors is more prevalent than bullying by colleagues, can be confirmed.This has previously been substantiated by the research of Hoel and Cooper (2000), Lutgen-Sandvik et al. (2007) and Namie and Namie (2003).
Bullying by superiors has a greater influence than bullying by colleagues on the targets leaving the organisation.
The significance of the correlations confirmed hypotheses three and four, namely that workplace bullying by superiors or colleagues has negative relationships with all the subfacets of POS (role clarity, job information, participation in decision-making, colleague support and supervisory relationships).This implies that when bullying increases, POS will decrease.These findings are consistent with recent studies suggesting that the employees affected by bullying in the workplace experience a lack of role clarity (Agervold, 2007;Hauge, Skogstad & Einarsen, 2010), important information will be kept from them on purpose (Johnson & Rea, 2009), they will have less control over decisions being made in the organisation (Notelaers, De Witte & Einarsen, 2010;Quine, 2001), they will experience less support from colleagues and supervisors (Notelaers & De Witte, 2003;Tsuno, Kawakami, Inoue & Abe, 2010).Supervisory relationships correlated with a large effect in comparison with the other sub-facets, which indicates the prevalence of poor supervisory relationships in a bullying culture as found in previous research.
Both bullying by superiors and bullying by colleagues correlated positively with turnover intention (providing support for hypotheses four and five), indicating that when the bullying by either superiors or colleagues increases, the target's intention to leave the organisation will also increase.This confirms the results of several previous studies that bullying icreases an individual's intention to leave the organisation (i.e.Berthelsen, Skogstad, Lau & Einarsen, 2011;Djurkovic et al., 2004;Hoel & Cooper, 2000).Bullying by superiors had a slightly higher correlation with turnover intention than bullying by colleagues, indicating that the influence of bullying by superiors on the intention to leave the organisation is greater than the influence of bullying by colleagues.
The POS in the organisation established its significance in the relationship with turnover intention, with a negative association, confirming hypothesis six.This implies that in the presence of the sub-facets of POS (role clarity, job information, participation in decision-making, colleague support and supervisory relationships) the turnover intention in the organisation will decrease, leading to targets not considering leaving the organisation.This result confirms the results of previous studies suggesting that turnover intention is influenced by the extent to which employees experience role clarity (Bhuian, Menguc & Borsboom, 2005;Hwang & Chang, 2009), have sufficient job information (Lambert, 2006), participate in the decision-making process (Knudsen, Ducharme & Roman, 2009), and are supported by their colleagues (Brough & Frame, 2004;Kim & Stoner, 2008) and supervisor (Lambert, 2006;Maertz et al., 2007).
Hypothesis seven stated that POS (role clarity, job information, participation in decision-making, colleague support and supervisory relationships) plays a moderating role in the relationship between workplace bullying and turnover intention.This hypothesis is partially confirmed.According to the results, POS (role clarity, participation in decisionmaking and supervisory relationships) acted as moderator in the relationship between workplace bullying by superiors and turnover intention.However, the results indicate that the POS variables job information and colleague support did not moderate the relationship between workplace bullying and turnover intention.Similarly, no moderation was found with bullying by colleagues.A study by Djurkovic et al. (2004) yielded similar findings of the moderating role of POS in the relationship between bullying and turnover intention, however, in this study the definition used to define POS differed slightly from the definition used in this study.
As a result, when superiors bully their subordinates and the targets have a clear understanding of what their roles in the organisation are, these victims will be more inclined to consider not leaving the organisation.Similarly, when they have a say in the decisions of the organisation the targets will also remain.Furthermore, when employees are bullied by superiors, but have excellent relationships with other superiors, then this will also counteract the effects of workplace bullying on the intention to leave the organisation.There were no sub-facets of POS acting as moderators in the relationship between bullying by colleagues and turnover intention.
To conclude, the primary objective of this study was to determine whether POS moderates the relationship between

TABLE 1 :
Characteristics of the participants.

TABLE 2 :
Descriptive statistics, Cronbach alpha coefficients and correlation coefficients of the SAEHWS.

TABLE 3 :
Correlation coefficients between workplace bullying (superiors and colleagues), POS and turnover intention.

TABLE 4 :
Hierarchical regression analysis (role clarity) with turnover intention as dependent variable.statistical significance; F, F-statistic; R, square root of R-square (correlation between the observed and predicted values of the dependent variable); R 2 , proportion of variance in the dependent variable explained by the independent variables; ΔR 2 , change in percentage variance explained by next step in model.*, p < 0.05 FIGURE 1: Interaction of bullying by superiors with role clarity in relation to turnover intention.

TABLE 5 :
Hierarchical regression analysis (participation in decision-making) with turnover intention as dependent variable.F, F-statistic; R, square root of R-square (correlation between the observed and predicted values of the dependent variable); R 2 , proportion of variance in the dependent variable explained by the independent variables; ΔR 2 , change in percentage variance explained by next step in model.*, p < 0.05 Participation FIGURE 2: Interaction of bullying by superiors with participation in decisionmaking in relation to turnover intention.

TABLE 6 :
Hierarchical regression analysis (supervisory relationship) with turnover intention as dependent variable., standard error; t, t-statistic; p, statistical significance; F, F-statistic; R, square root of R-square (correlation between the observed and predicted values of the dependent variable); R 2 , proportion of variance in the dependent variable explained by the independent variables; ΔR 2 , change in percentage variance explained by next step in model.*,p<0.05workplace bullying and turnover intention.The results of this study confirm this moderating effect, suggesting that when bullied by supervisors, a lack of POS (more specifically role clarity, participation in decision-making and supportive supervisory relationships) will increase bullying victims' propensity to leave the organisation.This intensifies the need for organisations to provide a supportive environment for their employees, especially those being targeted by bullies.Limitations in this study could be addressed in future research.The first limitation is that longitudinal research can compile more profound evidence of what is being researched, which was not the case with this research study.Another limitation is that this study focused on many sectors, which makes the counteraction strategies less specific to any particular sector.A further setback was that there was no satisfactory evidence to prove that POS can act as a moderator in the case of bullying by colleagues, and this could still be an important element in South African research.The findings of this research lay the foundation for further research with regard to workplace bullying in the South African context.Whereas this study gave a broad overview of the manifestation of workplace bullying in many sectors in South Africa, more in-depth research is necessary in order to provide a framework for the development of preventative interventions or support programmes for those individuals targeted with bullying.Longitudinal studies would allow researchers to draw accurate conclusions regarding the development of turnover intentions of individuals being bullied at work.Also, more information is needed regarding other organisational factors predicted by bullying in the workplace if a comprehensive model of workplace bullying is to be developed.Workplace bullying is a silent epidemic(McAvoy & Murtagh, 2003)and will remain so if no counteraction is encouraged through research and implemented in practice.
SEFIGURE 3: Interaction of bullying by superiors with supervisory relationships in relation to turnover intention.