• 1. Introduction

  It is inevitable in workplaces for a worker’s performance to be evaluated by supervisors, co-workers and subordinates. Others’ evaluation of work can elicit feelings of shame and pride. When we have not performed well, particularly when the person doing the evaluation is important to us, we are vulnerable to feelings of shame. When we have performed well, we are likely to feel pride. The management of shame and pride matters in workplace relations.
  Shame and pride have been conceptualised as conjugate emotions in social relationships (Nathanson, 1992; Scheff & Retzinger, 1991): shame is related to alienation, pride to social solidarity (p. xix). Other work, however, differentiates pro-social and anti-social forms of shame and of pride (Leach & Cidam, 2015; Tangney, 1990; Webb, 2003). Pride that is authentic involves self-respect with realisation of limitations and others’ contributions. As Webb (2003) puts it, authentic pride is tinged with humility and avoids hubris. Pride that manifests as hubris is associated with behavioural problems and is destructive of interpersonal relationships (Baumeister, 2001; Tangney, 1990). Such pride is a likely contributor to the power imbalance in bullying relationships.
  Similarly, shame has a negative and a positive face. On the negative side, shame that is unacknowledged is likely to be expressed as anger and defensiveness (Gilligan, 1997; Lewis, 1971; Tangney, 1990). Power imbalance arises with emotional domination, leading others to adopt placating or fearful responses. In contrast is the kind of shame that feels unpleasant but is positive in the sense that it is socially adaptive, signalling to us that we have caused harm and we need to make amends (Retzinger, 1996; referred to as guilt by Tangney, 1990). These ideas have been incorporated into the theory of shame and pride management through reintegration (Ahmed & Braithwaite, 2011; Ahmed et al., 2001; Braithwaite, Ahmed & Braithwaite, 2008), a revision of reintegrative shaming theory.

  1.1 The theory of shame and pride management through reintegration

  The essence of the revision of reintegrative shaming theory is that it adds a psychological dimension to the shaming process used by institutions to sanction wrongdoing or failure. From the perspective of an institution’s practices (for example, the criminal justice system or the education system or the child rearing system), shaming of a person over an act deemed unacceptable may be stigmatising (the whole person is at fault) or reintegrative (disapproval of the act, without rejection of the person). The insight of reintegrative shaming theory for institutional practices is that reintegrative disapproval has better outcomes for harm reduction and social cohesion than stigmatising disapproval (Braithwaite, 1989).
  But what an authority intends to do when sending a message of disapproval is not necessarily the same as how that message is received. How we make sense of the treatment we receive from others when we do something wrong is part of the revision of reintegrative shaming theory. The revised theory was supported by empirical findings emerging from the Reintegrative Shaming Experiments (RISE) from 1995 to 1999 in Canberra ACT and the Life at School Project, a study of school bullying from 1996 to 1999 in Canberra ACT (Ahmed et al., 2001).
  When an authority sanctions or rewards people, it is sending a message that potentially either affirms or threatens what Nathan Harris calls our ethical identity, our sense of being a worthy human being (Harris, 2007). Ethical identity represents our best self in terms of capability and character. Harris provides the first plank for revising the theory: our ethical identity is more likely to be ‘touched’ or awakened when those who matter most to us are part of the message of disapproval. We are able to discount the messaging from those who are not among our significant others, in particular, authority figures who we do not respect or admire.
  Second, once our ethical identity is evoked, so too is our sense of social well-being through the emotions of shame and pride. The revision of reintegrative shaming theory states that how we manage these emotions influences outcomes. We can manage these emotions well or poorly from the perspective of maintaining positive social relationships and abiding by social expectations (Ahmed et al., 2001; Ahmed & Braithwaite, 2004; Shin, 2005). When we do something that causes harm, depending on our nature and circumstance, shame can be acknowledged in the sense of admitting responsibility of causing harm and making amends (acknowledged shame), or it can be denied with shame being released as anger and blame towards others (displaced shame). Acknowledged shame is a bid to repair relationships. Displaced shame tends to make strained relationships worse. Managing these emotions well is helped by being in a ‘safe space’ where rituals of reintegration prevail over rituals of stigmatisation, but our temperament, cultural background and past experience also influence how we interpret signals from others and how we manage these emotions.
  The measurement of shame and pride management has been pioneered by Eliza Ahmed through the development of two inventories: The Management of Shame State – Shame Acknowledgement Shame Displacement (MOSS-SASD) and the Management of Pride State (MOPS) (Ahmed, 2001; Ahmed & Braithwaite, 2011). The inventories comprise sets of imaginary scenarios. Respondents place themselves in the scenarios and report their likely feelings and actions of shame and pride. The scenarios relate to being caught bullying (shame), failing to meet work goals (shame) or achieving their goals (pride). Responses cover blame, anger, remorse and repair for the MOSS-SASD, and superiority, dominance, social inclusion and humility for the MOPS. This work has taken place in Australia, Bangladesh and Korea and has demonstrated that personality, history, culture and context matter in how we manage shame and pride.

  1.1.1 Evidence linking shame and pride management to bullying

  Working with children in the Life at School Survey, Ahmed found that well-known predictors of bullying such as dislike for school and problems at school, authoritarian parenting, low empathy, low self-esteem and poor self-control were partially mediated in their relationship to bullying by levels of shame acknowledgement and shame displacement. Acknowledging shame characterised children who were less likely to be involved in bullying, while displacing shame characterised children who were bullying others (Ahmed et al., 2001; Ahmed & Braithwaite, 2004). After three years, Ahmed and Braithwaite (2012) found that children who had changed and moved into a bullying role were more likely to displace shame and judge themselves to be in a bullying tolerant school culture. Children who moved out of a bullying role were more likely to acknowledge shame and report their school as having controls against bullying. These studies demonstrated the importance of both the capacity to manage shame and the controls in place to stop children bullying at school. In a subsequent study using the Life at School Survey with 1,875 Bangladeshi students, Ahmed and Braithwaite (2006) found that children who perceived their parents as willing to forgive and reconcile with them over their wrongdoing were more likely to acknowledge shame and less likely to displace shame, which in turn was linked to less bullying. Personality, parenting, school practices and life experience are all linked with shame management among school children, and shame management is related to school bullying and desistence over time.
  Helene Shin (2005) used a modified version of the MOSS-SASD with Australian and Korean teachers to understand their capacity to manage shame when placed in scenarios as a perpetrator of workplace bullying. Shin examined the contextual, dispositional and cultural correlates of teachers’ shame management. Shame displacement was more common when teachers reported experiencing disrespectful treatment in their workplace and when they placed high value on competitiveness and achievement. Shame acknowledgement in response to the bullying scenarios was higher for those who valued collective well-being. These findings suggest that dealing effectively with wrongdoing requires responsiveness: Personalities and context both need to be understood before an action plan is implemented.
  At the time of revising reintegrative shaming theory, work on pride management was in its infancy. Tangney’s (1990) concept of pride proneness had been found to predict less school bullying (Ahmed & Braithwaite, 2004), but the effect was small compared to the much stronger predictor of shame management. The situation changed, however, when Ahmed’s conceptualisation of pride was used instead of that of Tangney. In workplaces pride in success could foment destructive social relationships. In a study of 824 Bangladeshi workers, Ahmed introduced the Management of Pride State (MOPS) inventory to be used alongside the MOSS-SASD (Ahmed & Braithwaite, 2011). This study demonstrated how positive and negative expressions of pride were related to workplace bullying. Confirmatory factor analysis revealed that MOPS had two factors. One represented expressions of triumph over others (narcissistic pride) which could be humiliating for others. The other represented expressions of shared triumph with others who had contributed towards success (humble pride). Employees involved in bullying had lower scores on humble pride (respecting self and others), and higher scores on narcissistic pride (feeling dominant and arrogant). They also expressed poor shame management with lower scores on shame acknowledgement (feeling shame/guilt, taking responsibility, making amends) and higher scores on shame displacement (hitting out at others, blaming others). Furthermore, shame displacement and narcissistic pride were more likely to occur when workplaces were perceived as being interpersonally disrespectful and not procedurally transparent. In these work conditions, shame acknowledgment was less likely to be expressed (Ahmed & Braithwaite, 2011; Braithwaite et al., 2008).

  1.2 Shame and pride management: a ‘red flag’ for resolving work bullying complaints?

  Like the workplace bullying literature reviewed briefly below, shame and pride management research reveals a confluence of forces leading to bullying encounters, some are contextual, some are cultural, some are normative, some are interpersonal and some are personal, defined by history and personality. The sheer complexity of the bullying problem in workplaces makes restorative justice and restorative practice interventions attractive options (Burford, Braithwaite & Braithwaite, 2019). The popularity of such interventions is evident in an exponential growth in the spheres of human activity in which restorative practice in social interactions has been advocated (see overviews in Burford et al., 2019; Gavrielides, 2019; Llewellyn & Morrison, 2018). Amidst this expanding social movement of restorative practices, it is timely to attend to evidence of likely boundaries around the appropriate use of restorative justice in dealing with cases of workplace bullying.
  The concepts of shame and pride management and the body of empirical findings around them in connection to bullying raise concerns that restorative justice may not always be the best way of handling workplace problems of this nature. If someone is not prepared to contain their narcissistic pride and if they are locked into displacing shame, a restorative justice intervention may cause more harm than it seeks to resolve. By the same token, shame and pride management are conceived as states (not fixed personality traits) that can and do change over time (Ahmed & Braithwaite, 2012). Therefore, it is not the case that a restorative intervention is never appropriate, just that in particular circumstances, it is not appropriate at a particular time. Practitioners make such judgments all the time.
  The contribution of this article is to identify specific ‘red flags’ that can assist in decision making about whether a workplace bullying complaint should be handled through a restorative justice meeting or not. While narcissistic pride and displaced shame are two red flags to reflect upon critically, a third possible red flag is social connectedness, or lack thereof. It is possible that people who shun connectedness to others are disinterested in managing shame and pride and are careless in dealing with others, resulting in bullying behaviour in work settings. Their low trust may mean they find it difficult to acknowledge shame (even in a reintegrative setting) and may persist in denying responsibility or blaming others for problems that occur in the workplace. A similar logic applies to pride. Being socially disconnected may mean seeing their accomplishments as theirs and theirs alone, leading to narcissistic expressions of pride. In other words, the relationships we have been observing between socially dysfunctional shame and pride management on one hand and bullying on the other may be an artefact of a personal preference to maintain social distance from others.
  A fundamental premise of restorative justice and restorative practice is that connectedness is an important social value. Social connectedness has benefits for health and well-being (Helliwell & Putnam, 2004). But some people shun connectedness, preferring to preserve their social distance from others. Braithwaite et al. (2013) found that those in the general public who refused to participate in a restorative justice conference were more likely to be low in trust and have little time for notions of forgiveness and rehabilitation. Shunning connectedness would be an important personal impediment to using restorative justice for workplace bullying.
  The key question addressed in this article, therefore, is whether we have identified three red flags cautioning against the use of restorative justice meetings to resolve bullying problems or one more fundamental red flag. Is the relationship of shame and pride management to bullying a consequence of being socially disconnected (meaning we have one red flag), or is the problem of poor shame and pride management something that is separate from low levels of trust and connection with others (meaning we have three red flags)? The answer to this question shapes the discussion of whether it is appropriate to use a restorative justice meeting to manage a particular case of bullying, and if so, how preparations for such a meeting should proceed.
  At this point, it is important to emphasise that whether or not a restorative justice meeting should be held to address workplace bullying depends not only on the personal circumstances of perpetrators and targets, but also on macro policy interventions. Restorative practice interventions for workplace culture change to deal with bullying and harassment are now in widespread use (for example, respectful relationships programmes),¹xSee, for example, Australian Public Service Commission. (2011). Respect: promoting a culture free from harassment and bullying in the APS. 4th ed. Canberra, ACT: Australian Public Service Commission. sometimes with allegiances to different academic literatures. Of particular importance is the mainstream workplace bullying literature. This literature is briefly reviewed below to demonstrate how it is compatible with, and in some ways has developed in parallel with research on restorative justice and practice.

  1.3 The mainstream workplace bullying literature and connections to restorative practices

  Bullying is widely recognised as a relational problem. A well accepted definition is: Bullying involves persistent, offensive, abusive, or intimidating behaviour that makes the target feel threatened, humiliated, stressed or unsafe at work (Di Martino, Hoel & Cooper, 2003). A substantial body of empirical research also shows the harm of bullying, ranging from low job satisfaction, commitment and loss of confidence through depression, anxiety, mental and physical disorders, and, in some cases, work-related suicide (Dollard, Dormann, Tuckey & Escart, 2017; Mikkelsen & Einarsen, 2002; Potter, Dollard & Tuckey, 2016; Quine, 1999; Sansone & Sansone, 2015). Bullying is a relational problem in which the priority is to provide healing for victims and desistance among bullies, along with awareness of the serious consequences of their actions. As such, bullying lends itself to restorative justice resolution.
  Explanations of workplace bullying can be found in the characteristics of individual bullies (Sheehan, 1999; Zapf & Einrasen, 2003), the characteristics of their targets (Coyne, Seigne & Randall, 2000; Matthiesen & Einarsen, 2001), workplace relationships (Einarsen, 1999) and workplace culture (Agervold & Mikkelsen, 2004; Leymann, 1996). Perpetrators have been linked with personal qualities such as impulsivity, emotional reactivity, cynicism, low tolerance for ambiguity and aggressiveness (e.g. Matthiesen & Einarsen, 2007). Victims on the other hand are more likely to have low self-esteem, poor social competence, neuroticism and to exhibit negative affectivity more commonly than most (Balducci, Fraccaroli & Schaufeli, 2011; Matthiesen & Einarsen, 2007). Restorative justice has been widely used in cases where there is vulnerability, domination and disempowerment. Restorative justice has been used successfully in such circumstances to restore dignity and empower victims (Pennell & Burford, 2000; Strang, 2002).
  Bullying episodes are not only triggered by individuals, they are triggered by the structure of workplaces. Workplaces characterised by role ambiguity, high work demands, interpersonal conflict and tyrannical or laissez-faire leadership (Hauge, Skogstad & Einarsen, 2007; Skogstad, Einarsen, Torsheim, Aasland & Hetland, 2007) are more likely to have a culture of bullying. Communication openness, a supportive work environment and providing bullied employees with recourse within the workplace have been found to reduce bullying and the problems it poses for targets (Bilgel, Aytac & Bayram, 2006; Daniel, 2004; Oluremi, 2007). Culture change programmes promoting respectful relations have grown out of this literature.
  Sometimes, however, management pays lip service to respectful relations and tolerates bullying when it suits production targets (Agervold & Mikkelsen, 2004; Ferris, 2004). In such cases, anti-bullying laws requiring workplaces to promote a socially responsible work culture and be responsive to bullying complaints may set standards that nudge management in the direction of change. Enforcement, however, can be difficult with lack of evidence and inconsistent accounts of events (Ferris, 2004). Problems are compounded because perpetrators of bullying are often victims of bullying: Provocation, retaliation and self-defence are common workplace bullying narratives to defend against censure by those with authority (Braithwaite, 2013; House of Representatives Standing Committee on Education and Employment, 2012). Bully victims are most vulnerable to illness and adjustment problems (Coyne, Craig & Chong, 2004). The complexity of these cases is an argument for restorative justice in so far as the approach emphasises healing, problem solving and future well-being for all parties, as opposed to blame and punishment.
  Individual, structural, managerial and legal forces work in concert to confuse the issue of workplace bullying. There is unlikely to be a simple cause or a simple solution. But if individuals do not reach a point collectively of acknowledging that bullying actions have occurred and are unacceptable, subsequent interventions are likely to fail, thereby putting organisations at risk of a poor psychosocial safety climate (Potter et al., 2016). It is the importance of the acknowledgement of bullying and the harm that it causes at all levels of the organisation that has made interventions for restorative practices a particularly attractive option for preventing or addressing a bullying workplace culture (see Liebmann, 2016 on the ambition of restorative cities). Against this background, a restorative justice meeting to deal with a workplace bullying incident makes good sense – in theory. The red flag relates to the micro level decision of whether those involved in the bullying are psychologically able to benefit from the encounter. More specifically, do they have capacity to manage shame and pride in a socially productive way? More fundamentally, do they value social connectedness enough to benefit from the experience?

  1.4 Objectives

  The present study has two specific objectives that address how restorative justice might align with the psychology of individuals involved in workplace bullying. The first objective is to determine the relative contributions of shame management and pride management to the prediction of bullying and victimisation. Previous research which established these relationships in the workplace was based in Bangladesh (Ahmed & Braithwaite, 2011). Bangladesh is an Islamic society. Muslim employees, sensitised through religious teaching to vaunting pride as a vice, are likely to associate hubris with bullying. This study serves the purpose of replicating the effects of pride management as well as shame management in a western setting. It is hypothesised that shame displacement over task failure and narcissistic pride over task success will be individual risk factors for bullying others. In contrast, shame acknowledgment over task failure and humble pride over task success will reduce likelihood of engaging in bullying.
  Predictions around victimisation are more difficult because how organisations respond to bullying impacts the way in which victims of bullying manage both shame and pride. For example, admitting wrongdoing and humility may make people targets (victims) in workplaces that tolerate bullying, particularly if their bullying peers interpret these responses as vulnerability and openness to domination.
  The second objective of this article is to determine whether shame and pride management uniquely contribute to the prediction of bullying and victimisation after controlling for the degree to which people value social connectedness to others. If the proposed relationship between shame and pride management capacities and bullying can be explained by a more generalised conception of connectedness to others, addressing shame and pride management directly may have limited effectiveness. Low social connectedness (low trust) may mean resistance to learning shame and pride management skills that lessen the likelihood of bullying incidents. More fundamentally, low social connectedness may mean refusal to take part in any associated intervention in the first place (Braithwaite et al., 2013).

• 2. Method

  2.1 Sample

  The Australian sample for ‘A cross national comparative study: Australian and Japanese Attitudes to Crime’ comprised 1,967 randomly selected respondents who replied to a postal survey questionnaire (response rate of 36.1 per cent; for further details see Huang, Braithwaite, Tsutomi, Hosoi & Braithwaite, 2012). Of this sample, 1,045 responded to the module asking about workplace bullying. This module targeted those who had been in the workforce in the past twelve months and is the sample used for this analysis.
  The survey sample was drawn from the Australian electoral roll. Voting is compulsory in Australia so the electoral roll provides reasonably good coverage for sample selection of the Australian population over eighteen years of age. Proportional sampling was applied according to the population of each state and territory. Sampling frames delivered as good a cross-section of the population as possible across Australian states and territories. The demographic profile of the sample is provided in Appendix 1. Just over half the respondents were female, average age was 48 years and almost half had a secondary school or diploma qualification, with just over a quarter leaving after junior high school and just under a quarter gaining a higher education qualification.
  The questionnaire was completed by the person who self-nominated as the head of household, and took on average between 30 and 40 minutes to complete. Questionnaires were returned by mail with postage pre-paid.

  2.2 Measures

  Gender was coded as male (1), female (0). Age was assessed in years. Personal income was measured in units of 1,000 dollars per year. Education was coded in terms of primary or junior secondary education (1), senior secondary education, certificate, diploma or advanced diploma qualifications (2), and bachelor, graduate diploma or postgraduate degrees (3).
  The predictor variables for this study were three measures of social connectedness, the MOSS-SASD and the MOPS (to measure shame and pride management). The outcome variables were workplace bullying and victimisation, measured with a modified version of Quine’s (1999) scales. All of these were continuous variables scored (or re-scored) such that higher scores indicated more of the characteristic being measured.
  The three social connectedness variables were: (a) a single item representing belief in the importance of forgiving offenders; (b) a single item representing belief in the importance of rehabilitating offenders; and (c) a multi-item scale representing distrust in others, particularly those who can provide professional support.
  Separate Confirmatory Factor Analyses (CFA) were used to validate the multi-item scales used to test the theoretically proposed model.²xWe used the following indices to assess the fit of the model to the data: the chi-square (χ²) statistic and the related degrees of freedom (df), ratio of χ² and df, the adjusted goodness of fit index (AGFI), the comparative fit index (CFI), the Tucker-Lewis index (TLI), the root mean square error of approximation (RMSEA), and the standardised root mean square residual (SRMR). In general, a χ²/df value <3.00 and AGFI, CFI and TLI values >.90 indicate an acceptable fit (with values >.95 being ideal (Brown, 2006). Further, RMSEA and SRMR values ≤.08 indicate a reasonable fit to the data, whereas values ≤.05 indicate excellent fit (Hu & Bentler, 1999). Note that the χ² statistic is very sensitive to sample size and hence is no longer relied upon as a basis for acceptance or rejection of a model (Schermelleh-Engel, Moosbrugger & Müller, 2003; Vandenberg, 2006). As a result, the use of multiple fit indexes has developed to provide a more holistic view of goodness of fit, taking account not only of sample size but also of model complexity and other relevant issues of the study such as multivariate normality. The seven variables measured through multi-item scales were: Distrust (representing poor social connectedness), shame acknowledgment and shame displacement (representing shame management), narcissistic pride and humble pride (representing pride management), workplace bullying and workplace victimisation. The items and scales and their development are discussed in detail in Appendix 1. Descriptive statistics and Cronbach alpha reliability coefficients are also in Appendix 1.

  2.3 Data analysis overview

  Zero-order correlational analyses between the outcome variables of bullying and victimisation, and the predictor variables of shame and pride management, and social connectedness revealed significant correlations linking social connectedness to shame and pride management and to bullying and victimisation (see Appendix 2). The interrelationships point to the need for multiple regression analysis to ascertain the effect of each variable above and beyond the effects of other variables.
  A set of regression models was tested. The purpose of the models was to measure the effect of the social connectedness variables after controlling for background factors, the effect of the shame and pride management variables after controlling for background factors, and the effect of social connectedness, pride and shame management in combination.
  The demographic variables that were controlled in the analyses were sex, age and status. Some studies report no significant demographic differences (e.g. Lee & Brotheridge, 2006), other studies do (e.g. Braithwaite et al., 2008; Einarsen, Hoel, Zapf & Cooper, 2011; Keuskamp, Ziersch, Baum & Lamontagne, 2012; Potter et al., 2016). Education level and personal income were used as proxies for status. Because of the very high correlation between bullying others and being bullied by others (r = 0.50, p < 0.001), regression models predicting bullying controlled for being a victim, and regression models predicting being a victim controlled for bullying.
  The order of entry was the same for bullying and victimisation. In Model 1, demographic variables were entered to examine their contribution to bullying and victimisation. Model 2 featured demographics + the bullying control variable (or the victimisation control variable). Model 3 tested the effects of shame and pride management beyond those of the demographics + the bullying control variable (or the victimisation control variable). Model 4 tested the effects of social connectedness beyond those of the demographics + the bullying control variable (or the victimisation control variable). Model 5 tested the full Model containing demographics + the bullying control variable (or the victimisation control variable) + shame and pride management + social connectedness. This set of analyses allowed us to test for the effects of shame and pride management with and without controlling for the social connectedness variables.

• 3. Results

  3.1 Predicting bullying

  The results for the regression models in which bullying was regressed on different sets of predictor variables are presented in Table 1. All models were statistically significant as shown by the F statistics underneath each model in Table 1. Consistently across the models in Table 1, gender was a significant predictor of bullying, with males more often reporting bullying than females (beta ranged from 0.11 to 0.18, p < 0.001 across models).

  Table 1 Standardised beta coefficients for five models predicting bullying (n = 1,045)

  Variables	Model 1 Demog	Model 2 Demog + Status	Model 3 Demog + Status + Shame & Pride	Model 4 Demog + Status + Connectedness	Model 5 Full Model
  Gender (male)	0.18***	0.16***	0.12***	0.13***	0.11***
  Age	–0.07*	–0.01 (ns)	0.03 (ns)	–0.01 (ns)	0.02 (ns)
  Education	–0.04 (ns)	–0.03 (ns)	–0.04 (ns)	–0.01 (ns)	–0.03 (ns)
  Personal income	0.04 (ns)	0.02 (ns)	0.00 (ns)	0.02 (ns)	0.00 (ns)
  Victimisation	-	0.49***	0.43***	0.47***	0.42***
  Forgive offender			-	–0.02 (ns)	0.00 (ns)
  See offender rehabilitate	-	-	-	–0.07*	–0.06*
  Distrust in professionals	-	-	-	0.08**	0.04 (ns)
  Shame acknowledgement	-	-	–0.03 (ns)	-	–0.02 (ns)
  Shame displacement	-	-	0.05 (ns)	-	0.05 (ns)
  Humble pride	-	-	–0.07**	-	–0.06*
  Narcissistic pride	-	-	0.24***	-	0.23***
  Adj R square	0.035	0.272	0.337	0.283	0.341
  F value	10.54***	78.88***	59.93***	52.43***	45.99***

  ***p < 0.001, **p < 0.01, *p < 0.05

  
  The shame and pride variables added more explanatory variance above and beyond the control variables (from 27 per cent to 34 per cent, see Model 3) than social connectedness (from 27 per cent to 28 per cent, see Model 4). Of most importance was narcissistic pride followed by humble pride. Those who bullied others were more likely to express narcissistic pride (beta = 0.24, p < 0.001) and less likely to express humble pride (beta = –0.07, p < 0.05). Shame acknowledgement and displacement were not important as predictors of bullying when the pride variables were also in the regression model. Of note was the finding that without the pride management variables in the equation, shame displacement was positively related to bullying, but shame shame acknowledgement was not.
  The social connectedness variables added only a small amount of explanatory power above and beyond the control variables (see Model 4). Believing that it was important to see offenders rehabilitated was less likely to be endorsed by those who bullied (beta = –0.07, p < 0.05). Those who bullied were more likely to express distrust in others (beta = 0.08, p < 0.01), particularly professionals who could potentially help resolve bullying problems.
  When all predictors were entered in the final model, being male (beta = 0.11, p < 0.001), rejection of the importance of rehabilitating offenders (beta = –0.06, p < 0.05), rejection of humble pride (beta = –0.06, p < 0.05) and endorsement of narcissistic pride (beta = 0.23, p < 0.001) remained as significant predictors of bullying.
  The beta coefficient for distrust became non-significant in the final regression. One possible explanation is that it shared variance with a number of other stronger predictors. From the bivariate correlations (see Appendix 2), distrust was correlated modestly with a host of variables besides bullying: Distrust was higher for men (r = 0.20, p < 0.001), for those with less education (r = –0.20, p < 0.001), for those who did not place importance on forgiveness (r = –0.18, p < 0.001) or rehabilitation (r = –0.13, p < 0.001), for those who did not practice humble pride (r = –0.16, p < 0.001) and for those who were prone to narcissistic pride (r = 0.16, p < 0.001).

  3.2 Predicting victimisation

  The results for the regression models in which victimisation was regressed on different sets of predictor variables are presented in Table 2. All models were statistically significant as shown by the F statistics underneath each model in Table 2. Consistently across the models in Table 2, age was a significant predictor of victimisation, with younger people more often reporting being the victim of bullying than older people (beta ranged from 0.07 to 0.11, p < 0.05 across models).

  Table 2 Standardised beta coefficients for five models predicting victimisation (n = 1,045)

  Variables	Model 1 Demog	Model 2 Demog + Status	Model 3 Demog + Status + Shame & Pride	Model 4 Demog + Status + Connectedness	Model 5 Full Model
  Gender (male)	0.04 (ns)	–0.05 (ns)	–0.03 (ns)	–0.06*	–0.05 (ns)
  Age	–0.11***	–0.08**	–0.07*	–0.08**	–0.08**
  Education	–0.04 (ns)	–0.02 (ns)	–0.03 (ns)	0.01 (ns)	0.00 (ns)
  Personal income	0.04 (ns)	0.02 (ns)	0.02 (ns)	0.03 (ns)	0.03 (ns)
  Bullying	-	0.50***	0.46***	0.48***	0.45***
  Forgive offender	-	-	-	–0.02 (ns)	–0.04 (ns)
  See offender rehabilitate	-	-	-	0.06	0.05 (ns)
  Distrust in professionals	-	-	-	0.12***	0.12***
  Shame acknowledgement	-	-	0.10***	-	0.10***
  Shame displacement	-	-	0.11***	-	0.11***
  Humble pride	-	-	0.04 (ns)	-	0.06*
  Narcissistic pride	-	-	0.01 (ns)	-	0.00
  Adj R square	0.011	0.253	0.282	0.267	0.296
  F value	3.86**	71.81***	46.51***	48.53***	37.55***

  ***p < 0.001, **p < 0.01, *p < 0.05

  
  The shame and pride variables added modest explanatory variance above and beyond the control variables (from 25 per cent to 28 per cent, see Model 3) as did the social connectedness variables (from 25 per cent to 27 per cent, see Model 4). In the prediction of victimisation, the management of shame, both acknowledged and displaced, was more important than the management of pride. The pride management variables lost explanatory power in the presence of the shame management variables. Those who were victimised were more likely to express acknowledged shame (beta = 0.10, p < 0.001) and displaced shame (beta = 0.11, p < 0.001).
  Of the social connectedness variables (see Model 4), distrust was the only variable that predicted victimisation over and above the control variables. Those who had been victimised were more likely to distrust others (beta = 0.12, p < 0.001).
  When all predictors were entered in the final model, being younger (beta = –0.08, p < 0.01), being distrustful of others (beta = 0.12, p < 0.001), and acknowledging shame and displacing shame (beta = 0.10 and 0.11 respectively, p < 0.001) remained as significant predictors of bullying. The beta coefficient for humble pride became significant in the final regression (beta = 0.06, p < 0.05), greater humble pride being associated with victimisation.
  In sum, bullying was primarily associated with pride management and victimisation was primarily associated with shame management. Shame and pride management remained significant predictors after the social connectedness variables were entered into the models. Those who bullied others were more likely to endorse narcissistic pride over humble pride. In addition, they could see little importance in rehabilitating offenders and showed signs of distrust. Those who were victimised were prone to acknowledge and displace shame. They were more likely to express distrust and endorsed humble pride in the workplace.

• 4. Discussion

  The purpose of this article has been twofold. First, do the relationships of shame and pride management with bullying and victimisation replicate findings from Bangladesh and offer further support for the revised theory of Shame and Pride Management through Reintegration (Ahmed et al., 2001; Ahmed & Braithwaite, 2011; Braithwaite et al., 2008)? The first objective was partially supported: pride and shame management are relevant factors in bullying incidents, with poor pride management associated with perpetrators and poor shame management associated with targets.
  The second purpose was to investigate whether the pride and shame management relationships with bullying and victimisation could be explained by the degree to which people want to restore or avoid social connections to others. In other words, do people who disconnect from others, and therefore possibly have little interest in learning social skills, manifest a set of anti-social behaviours that include poor shame and pride management skills on the one hand and involvement in bullying encounters on the other?
  The second proposition was resolved in the negative. There was no evidence that observed relationships between shame and pride management and bullying were an artefact of keeping one’s distance from other people. Shame and pride management and social disconnectedness were both related to bullying and victimisation, but their effects were largely independent of each other. Trust was low for both bullies and victims, regardless of shame and pride management skills, and bullies placed little importance on the more general principle of rehabilitation for offenders. The use of restorative justice depends on dealing with such impediments separately and possibly, sequentially. For example, if trust issues could be settled to the point of convincing the parties to take part in a restorative justice meeting, issues around shame and pride management could be addressed subsequently as part of discussions about behaviours that would help everyone feel safe and comfortable in the workplace.
  Not all the relationships between pride and shame management, and bullying and victimisation from the Bangladesh study (Ahmed & Braithwaite, 2011) held up in the Australian context. Being a bully involved poor pride management, that is, more narcissistic pride and less humble pride. Being a victim involved poor shame management in the form of displaced shame, as well as the more positive form of acknowledged shame and humble pride. Possibly this separation of pride management for understanding bullying and shame management for understanding victimisation has come about because controls were introduced for being a victim in the prediction of bullying and for being a bully in the prediction of victimisation. Pure bullying is associated with poor pride management, pure victimisation is associated with poor shame management, and those who are both bully and victim will more than likely have both poor pride and shame management skills (as occurred in the Bangladesh study).
  The pattern of being both a bully and a victim in the past twelve months is possibly a reflection of workplace norms where victims are encouraged to stand up for themselves, and respond to bullying with bullying. Bullying and reciprocation of bullying has been noted in the literature (e.g. Lee & Brotheridge, 2006). A feature of bullying cultures is that they tolerate bullying chains: Senior managers bully junior managers, junior managers bully supervisors, supervisors bully workers down the hierarchy (Vandekerckhove & Commers, 2003). The extent to which a workplace has effective policies to control bullying and policies around respectful relations should decrease the prevalence of bully victims. The present survey, however, does not allow for exploration of this issue because it does not provide data on workplace culture. This raises the important issue of the limitations of this research.

  4.1 Limitations

  This survey focused on the experiences of individuals in their communities and workplaces. Communities varied and workplaces varied. The survey was useful for understanding the beliefs and values of a broad cross-section of Australians regarding the harms they faced on the streets and at work, and their attitudes to these harms, in particular, their readiness to embrace restorative practices and restorative justice conferencing. This methodology departs from mainstream workplace bullying research in focusing on individuals as citizens and averaging across the various workplaces in which Australians are employed, and the conditions under which they are employed.
  The redeeming feature of these data is that they were collected at a time when workplace bullying was not a salient issue in the public eye and there was not a public narrative about its causes and effects. These data, therefore, offer valuable insights into how individuals who experience workplace bullying in a variety of settings make sense of their own experiences and respond to them, without the influence of expert commentary and social media. That said, it would be interesting to know how pride and shame management interact with work culture. Undoubtedly work culture matters, and having such data available would inform what is currently a supposition. That interventions supporting culture change towards more restorative practice would provide a safe harbour for restorative justice meetings to address shame and pride management, should those involved in bullying incidents be willing to take part.
  A second limitation of this study is the cross-sectional nature of the data which means that it is impossible to tease out causal sequences among the key variables – social connectedness, shame and pride management and bullying. Previous studies have used structural equation models to demonstrate the plausibility of shame and pride management as mediating variables (Ahmed & Braithwaite, 2011; Braithwaite et al., 2008), but longitudinal data is required to fully establish causal processes. What is known from longitudinal data is that shame management skills can be learnt by children. We can only hope that adults also can learn better strategies for managing shame and pride in the workplace, given the opportunity.

  4.2 Implications and conclusion

  This article has sought to provide evidence of where boundaries might lie in the use of restorative justice for dealing with incidents of workplace bullying. Practitioners are well versed in spotting these boundaries. Furthermore, both theory and empirical findings raise questions around the suitability of restorative justice interventions in workplaces to repair the harm of bullying and harassment, with much of this debate revolving around power. Skilled practitioners have developed their capacities to manage the red flag of power and domination. But this expertise is not within the public’s reach. The public is left to trust. The contribution of this article is to explain how these workplace power differentials can become so damaging. This article grounds these justifiable concerns about power in measurable, psychologically meaningful and understandable concepts; concepts that can be openly discussed with workers, managers, families and the public at large.
  Shame and pride are human emotions with which we are all familiar. If we better understand harmful expressions of these emotions as well as socially productive expressions we are better positioned to rein in others’ excesses as well as our own. We are also better positioned to challenge workplace practices that incentivise poor shame management and poor pride management. There are many workplaces where workers are afraid to admit mistakes for fear of losing their jobs or pay, and find themselves displacing shame rather than acknowledging it. There are many workplaces that set performance targets and bonuses for staff that encourage narcissistic pride and dismiss humble pride as an expression of outmoded modesty.
  This article set out to consider the question of whether there were three red flags for those contemplating using restorative justice meetings to resolve a workplace bullying incident (displaced shame, narcissistic pride and low connectedness) or just one fundamental condition (low connectedness). The answer is that all three must be considered. Low connectedness is a problem that practitioners take in their stride, seeking ways to build trust so that healing conversations can take place. Displaced shame and narcissistic pride will be familiar too, but perhaps not in such a theoretically coherent way. To desist from bullying, pride and shame management need to be directly addressed. And pride and shame management are hardest to control when levels of arousal are high. Perpetrators and targets need to be willing to look at themselves and how their pride and shame management capacities might need to change, particularly under pressure, if they are to be good work colleagues.
  No one should be asked to meet these conditions unless the workplace environment is one in which it is safe to do so. Macro-level restorative interventions to change workplace culture provide a necessary backdrop to the kind of restorative justice meeting proposed here. This is an important avenue for further research to understand the interplay between macro restorative interventions and micro restorative interventions that seek to improve individuals’ capacities to manage pride and shame.
  In conclusion, there will be those who are unwilling or unable to reflect on their pride and shame management skills. Or are unwilling to tolerate a process that requires greater interpersonal connectedness. And there will be organisations that have policies that fuel rather than defuse poor pride and shame management among their workers and do nothing to build workers’ trust in them. With this in mind, this article locates itself within a tradition that considers restorative justice within a framework of legal pluralism (Aertsen, Daems & Robert, 2006; Walgrave, 2008). Many warn of western legal systems crowding out or distorting the best of what restorative justice can offer communities. But restorative justice can be equally discredited by overpromising and not delivering because it has failed to recognise the boundaries of its effectiveness and failed to integrate with a raft of other measures for dealing with the harm that people do to each other – within families, within workplaces and to strangers (Burford et al., 2019). Understanding the workings of pride and shame management in workplaces is one step towards having increasingly open conversations about the limitations of restorative justice, without denying anyone the opportunity of using this process to repair relationships with others.

References

• Aertsen, I., Daems, T. & Robert, L. (eds.) (2006). Institutionalizing restorative justice. Cullompton: Willan Publishing.

• Agervold, M. & Mikkelsen, E.G. (2004). Relationships between bullying, psychosocial work environment and individual stress reactions. Work and Stress, 18(4), 336-351. doi:10.1080/02678370412331319794.

• Ahmed, E. (2001). Shame management: regulating bullying. In E. Ahmed, N. Harris, J. Braithwaite & V. Braithwaite (eds.), Shame management through reintegration (pp. 211-314). Cambridge: Cambridge University Press.

• Ahmed, E. & Braithwaite, V. (2004). ‘What? Me ashamed?’ Shame management and school bullying.Journal of Research in Crime and Delinquency, 41(3), 269-294. doi:10.1177/0022427804266547.

• Ahmed, E. & Braithwaite, V. (2006). Forgiveness, reconciliation, and shame: three key variables in reducing school bullying. Journal of Social Issues, 62(2), 347-370. doi:10.1111/j.1540-4560.2006.00454.x.

• Ahmed, E. & Braithwaite, J. (2011). Shame, pride and workplace bullying. In S. Karstedt, I. Loader & H. Strang (eds.), Emotions, crime and justice(pp. 55-79). Oxford: Hart Publishing.

• Ahmed, E. & Braithwaite, V. (2012). Learning to manage shame in school bullying: lessons for restorative justice interventions. Critical Criminology, Special Issue on Restorative Justice: Unraveling the Mystery, 20(1), 79-97.

• Ahmed, E., Harris, N., Braithwaite, J. & Braithwaite, V. (2001). Shame management through reintegration. Cambridge: Cambridge University Press.

• Balducci, C., Fraccaroli, F. & Schaufeli, W.B. (2011). Workplace bullying and its relation with work characteristics, personality, and post-traumatic stress symptoms: an integrated model. Anxiety, Stress & Coping: An International Journal, 24(5), 499-513. doi:10.1080/10615806.2011.555533.

• Baumeister, R.F. (2001). Violent pride. Scientific American, 284(4), 96-101. doi:10.1038/scientificamerican0401-96.

• Bilgel, N., Aytac, S. & Bayram, N. (2006). Bullying in Turkish white-collar workers. Occupational Medicine, 56(4), 226-231. doi:10.1093/occmed/kqj041.

• Braithwaite, J. (1989). Crime, shame and reintegration. New York: Cambridge University Press.

• Braithwaite, V. (2013). A multipronged approach to the regulation of workplace bullying (Regulatory Institutions Network Occasional Paper 20). Submission to the Parliamentary Inquiry into Workplace Bullying, House of Representatives, Parliament House. Australian National University, Canberra.

• Braithwaite, V., Ahmed, E. & Braithwaite, J. (2008). Workplace bullying and victimization: the influence of organizational context, shame and pride. International Journal of Organizational Behaviour, 13(2), 71-94.

• Braithwaite, V., Huang, H. & Reinhart, M. (2013). Love thy neighbour’s values, needs, and willingness to participate in restorative justice: a survey of Australian and Japanese victims and offenders. Restorative Justice: An International Journal, 1(1), 91-121. doi:10.5235/20504721.1.1.91.

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

• Burford, G., Braithwaite, J. & Braithwaite, V. (2019). Restorative and responsive human services. London: Routledge.

• Coyne, I., Craig, J. & Smith-Lee Chong, P. (2004). Workplace bullying in a group context. British Journal of Guidance and Counselling, 32(3), 301-317. doi:10.1080/03069880410001723530.

• Coyne, I., Seigne, E. & Randall, P. (2000). Predicting workplace victim status from personality. European Journal of Work and Organizational Psychology, 9(3), 335-349. doi:10.1080/135943200417957.

• Daniel, B. (2004). Workplace bullying: a communication perspective (Unpublished Ph.D.). The Florida State University, Florida.

• Di Martino, V., Hoel, H. & Cooper, C.L. (2003). Preventing violence and harassment in the workplace. Dublin: European Foundation for the Improvement of Living and Working Conditions.

• Dollard, M., Dormann, C., Tuckey, M.R. & Escart, J. (2017). Psychosocial safety climate (PSC) and enacted PSC for workplace bullying and psychological health problem reduction. European Journal of Work and Organizational Psychology, 26(6), 844-857. doi:10.1080/1359432x.2017.1380626.

• Einarsen, S. (1999). The nature and causes of bullying at work. International Journal of Manpower, 20, 16-27. doi:10.1108/01437729910268588.

• Einarsen, S., Hoel, H., Zapf, D. & Cooper, C.L. (eds.) (2011). Bullying and harassment in the workplace: developments in theory, research, and practice. Boca Raton: Taylor & Francis.

• Ferris, P. (2004). A preliminary typology of organizational response to allegations of workplace bullying: ‘See no evil, hear no evil, speak no evil’. British Journal of Guidance & Counselling, 32(3), 389-395. doi:10.1080/03069880410001723576.

• Gavrielides, T. (2019). Routledge international handbook of restorative justice. London: Routledge.

• Gilligan, J. (1997). Violence: reflections on a national epidemic. New York: Vintage Books.

• Harris, N. (2007). Shame, ethical identity and conformity: lessons from research on the psychology of social influence (Regulatory Institutions Network Occasional Series Paper No 11). Canberra: Australian National University.

• Hauge, L.H., 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. doi:10.1080/02678370701705810.

• Helliwell, J. & Putnam, R.D. (2004). The social context of well-being. Philosophical Transactions of the Royal Society (London), 359, 1435-1446. doi:10.1098/rstb.2004.1522.

• House of Representatives Standing Committee on Education and Employment. (2012). Inquiry into workplace bullying. Workplace bullying: we just want it to stop. October, Canberra.

• Hu, L.T. & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55. doi:10.1080/10705519909540118.

• Huang, H., Braithwaite, V., Tsutomi, H., Hosoi, Y. & Braithwaite, J. (2012). Social capital, rehabilitation, tradition: support for restorative justice in Japan and Australia. Asian Journal of Criminology, 7(4), 295-308. doi:10.1007/s11417-011-9111-1.

• Keuskamp, D., Ziersch, A.M., Baum, F.E. & Lamontagne, A.D. (2012). Workplace bullying a risk for permanent employees. Australian & New Zealand Journal of Public Health, 36(2), 116-119. doi:10.1111/j.1753-6405.2011.00780.x.

• Leach, C.W. & Cidam, A. (2015). When is shame linked to constructive approach orientation? A meta-analysis. Journal of Personality and Social Psychology, 109(6), 983-1002. doi:10.1037/pspa0000037.

• Lee, R.T. & Brotheridge, C.M. (2006). When prey turns predatory: workplace bullying as a predictor of counter aggression/bullying, coping, and well-being. European Journal of Work and Organizational Psychology, 15(3), 352-377. doi:10.1080/13594320600636531.

• Lewis, H.B. (1971). Shame and guilt in neurosis. New York: International Universities Press.

• Leymann, H. (1996). The content and development of mobbing at work. European Journal of Work and Organizational Psychology, 5, 165-184. doi:10.1080/13594329608414853.

• Liebmann, M. (2016, June 22-24). Building the restorative city. Powerpoint presentation for 9th International Conference of the European Forum for Restorative Justice, Leiden: The Netherlands.

• Llewellyn, J.J. & Morrison, B. (2018). Deepening the relational ecology of restorative justice. The International Journal of Restorative Justice, 1(3), 343-355. doi:10.5553/ijrj/258908912018001003001.

• Matthiesen, S.B. & Einarsen, S. (2001). MMPI-2 Configurations among victims of bullying at work. European Journal of Work and Organizational Psychology, 10(4), 467-484. doi:10.1080/13594320143000753.

• 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. doi:10.1891/088667007782793174.

• Mikkelsen, E.G. & Einarsen, S. (2002). Basic assumptions and post-traumatic stress among victims of bullying at work. European Journal of Work and Organizational Psychology, 11(1), 87-111. doi:10.1080/13594320143000861.

• Nathanson, D. (1992). Shame and pride: affect, sex, and the birth of the self. New York: Norton.

• Oluremi, B.A. (2007). Communication openness, conflict events and reactions to conflict in culturally diverse workgroups. Cross Cultural Management, 14(2), 105-124. doi:10.1108/13527600710745723.

• Pennell, J. & Burford, G. (2000). Family group decision making: protecting children and women. Child Welfare, 79(2), 131-158.

• Potter, R., Dollard, M.F. & Tuckey, M.R. (2016). Bullying & harassment in Australian workplaces: results from the Australian Workplace Barometer Project 2014/2015. Canberra: Safe Work Australia.

• Quine, L. (1999). Workplace bullying in NHS Community Trust: staff questionnaire survey. British Medical Journal, 318, 228-232. doi:10.1136/bmj.318.7178.228.

• Retzinger, S. (1996). Shame and the social bond. In D. Parker, R. Dalziel & I. Wright (eds.), Shame and the modern self (pp. 6-20). Victoria: Australian Scholarly Publishing.

• Sansone, R.A. & Sansone, L.A. (2015). Workplace bullying: a tale of adverse consequences. Innovations in Clinical Neuroscience, 12(1-2), 32-37.

• Scheff, T.J. & Retzinger, S.M. (1991). Emotions and violence: shame and rage in destructive conflicts. Lexington: Lexington Books/DC Heath & Company.

• Schermelleh-Engel, K., Moosbrugger, H. & Müller, H. (2003). Evaluating the fit of structural equation models: tests of significance and descriptive goodness of-fit measures. Methods of Psychological Research Online, 8(2), 23-74.

• Sheehan, M. (1999). Workplace bullying: responding with some emotional intelligence. International Journal of Manpower, 20(1/2), 57-66. doi:10.1108/01437729910268641.

• Shin, H.H. (2005). Institutional safe space and shame management in workplace bullying (Unpublished PhD Dissertation). https://openresearch-repository.anu.edu.au/handle/1885/48189. Canberra: Australian National University.

• Skogstad, A., Einarsen, S., Torsheim, T., Aasland, M.S. & Hetland, H. (2007). The destructiveness of laissez-faire leadership behaviour. Journal of Occupational Health Psychology, 12(1), 80-92.

• Strang, H. (2002). Repair or revenge. Victims and restorative justice. Oxford: Oxford University Press.

• Tangney, J. (1990). Assessing individual differences in proneness to shame and guilt: development of the self-conscious affect and attribution inventory. Journal of Personality and Social Psychology, 59(1), 102-111. doi:10.1037/0022-3514.59.1.102.

• Vandenberg, R.J. (2006). Statistical and methodological myths and urban legends. Organizational Research Methods, 9(2), 194-201.

• Vandekerckhove, W. & Ronald Commers, M.S. (2003). Downward workplace mobbing: a sign of the times? Journal of Business Ethics, 45(1-2), 41-50.

• Walgrave, L. (2008). Restorative justice, self-interest and responsible citizenship. Cullompton: Willan Publishing.

• Webb, T. (2003). Towards a mature shame culture: theoretical and practical tools for personal and social growth (Unpublished Ph.D. Dissertation). University of Western Sydney.

• Zapf, D. & Einarsen, S. (2003). Individual antecedents of bullying: victims and perpetrators. In S.E. Einarsen, H. Hoel, D. Zapf & C.L. Cooper (eds.), Bullying and emotional abuse in the workplace: International perspectives in research and practice (pp. 165-184). London: International Perspectives in Research and Practice, Taylor & Francis.

Appendix 1

• A.1 Demographic variables

  Based on previous studies (see Braithwaite et al., 2008; Einarsen, Hoel, Zapf & Cooper, 2011; Keuskamp, Ziersch, Baum & Lamontagne, 2012), four variables were used as controls in the regression analyses: respondent gender, age, level of educational attainment and personal annual income. All but gender were used as interval, continuous variables.
  Male respondents (scored 1) were 46 per cent of the sample (females coded 0). Mean age was 47.8 years; SD = 15.6). Level of educational attainment and personal income were used as proxy for socioeconomic status. Educational attainment was measured via a single item (nine-point scale) with 1 representing little or minimum formal schooling and 9 representing a postgraduate level of education. The responses were grouped into three categories to be used as a quasi-interval scale: from little schooling to Junior Secondary (Year 10) (29 per cent), Senior Secondary, Certificate, Diploma or Advanced Diploma qualifications (47 per cent), and Bachelor or Graduate Diploma or Postgraduate Degree qualifications (24 per cent). As for personal income, the median was A$25,000 (Mean = 31,902; SD = 30,959).

  A.1.1 Social connectedness variables

  1. In the context of surveying the needs of victims of crime, a single item measure for Forgiving asked respondents to indicate how important they believed it was ‘For the victim to be able to forgive the offender’. They responded on a five-point scale from ‘not at all important’ (1) through ‘neither important nor unimportant’ (3) to ‘very important’ (5) (Mean = 3.39; SD = 1.06).

  2. Using the same five-point scale, a single item measure for Seeing rehabilitation asked respondents to indicate how important they believed it was ‘For the victim to see the offender rehabilitated’ (Mean = 3.48; SD = 0.99).

  3. Distrust is a five-item scale that measures the degree to which individuals express lack of trust in (a) governmental agencies, (b) medical doctors, (c) lawyers, (d) professional counsellors and (e) people (‘Ultimately, I cannot trust human beings’). The context for measurement was ‘in general’, but all could potentially be relevant as sources of help in cases of bullying.

  A CFA was used to test the extent to which the items fit a one-factor model. The CFA showed that the one-factor model had an excellent fit to the data (Model fit: χ² = 2.31; df = 4; χ²/df = 0.68; AGFI = 0.999; CFI = 1.00; TLI = 1.0; RMSEA = 0.000; SRMR = 0.006). The standardised loadings were high for all items ranged between 0.63 and 0.89. There were no recommended modification indices.
  Accordingly, scale scores were formed through averaging responses to each item on a five-point rating scale from ‘strongly disagree’ (1) to ‘strongly agree’ (5) (Mean = 2.65; SD = 0.62; Cronbach’s alpha = 0.67).

  A.1.2 Management of shame and pride

  (1) Shame management was measured through the Management of Shame State-Shame Acknowledgment and Shame Displacement (MOSS-SASD) scales, a scenario-based self-report measure with strong psychometric properties (for details see Ahmed et al., 2001).
  In this study, the MOSS-SASD items were contextualised by using one of the most common incidents (threat to professional status) that occurs in workplaces. Respondents were asked:

  Suppose that you were required to solve an old and difficult problem at your workplace. You could not solve it successfully. How likely is it that you would feel the following: …

  Respondents were then presented with a list of ten shame-related reactions representing acknowledgement and displacement). Each was rated on a four-point rating scale (1 = not likely, 2 = may happen, 3 = likely, 4 = almost certain).
  The CFA indicated only a few modification indices which have been used to enhance the model fit. Upon inspection of the factor loadings, one item³x‘Pretend that nothing had happened’ representing shame displacement. loaded relatively low (0.22) and was deleted from the model. Deletion of the item improved the model significantly and confirmed that the latent structure of shame management is best represented by the two factor model (Model fit: χ² = 31.80; df = 13; p < 0.003; χ²/df = 2.44; AGFI = 0.988; CFI = 0.998; TLI = 0.992; RMSEA = 0.027; SRMR = 0.005). All remaining loadings were significant, positively linked to the corresponded latent factor, and substantial in size ranging from 0.39 to 0.94.
  Based on the above CFA as well as prior conceptualisation and analyses of other survey data, a six-item Shame Acknowledgement scale comprised the average of ratings for these items: (a) Feel ashamed of yourself; (b) Feel you had let down your co-workers; (c) Feel you had harmed your reputation; (d) Feel insecure about what others thought of you; (e) Feel that you were being noticed by everyone in a negative way; and (f) Regret that you could not solve the problem (Mean = 2.04; SD = 0.64; Cronbach’s alpha = 0.85).
  The Shame Displacement scale comprised the average of ratings for these three items: (a) Feel angry with your co-workers; (b) Feel that you were being unfairly treated by being given such a task; and (c) Feel like placing the blame somewhere else for not being able to solve the problem (Mean = 1.37; SD = 0.47; Cronbach’s alpha = 0.71).
  (2) Pride management was measured using a scenario based self-report measure – the Management of Pride State scales (MOPS; for details on its psychometric properties, see Ahmed and Braithwaite, 2011). The MOPS items were contextualised by using the same basic scenario presented for shame management with failure at a work task being replaced by success. The scenario represented the attribution of competence and enhancement of professional status. Respondents were asked:

  Suppose that you were required to solve an old and difficult problem at your workplace. You solved it successfully. How likely is it that you would feel the following: …

  Respondents were then presented a list of twenty pride-related reactions related to narcissistic and humble pride (a complete list of items are available on request) to be rated on a four-point rating scale (1 = not likely, 2 = may happen, 3 = likely, 4 = almost certain).
  The CFA results indicated that eight items had very low factor loadings (lower than 0.12). Hence, these items were deleted to enhance the model fit. Examining CFA modification indices suggested model fit could be improved by co-varying error terms in the model. Adding these error terms significantly improved the model fit (χ² = 38.46; df = 25; p < 0.14; χ²/df = 1.30; AGFI = 0.987; CFI = 0.997; TLI = 0.995; RMSEA = 0.016; SRMR = 0.010) suggesting that its latent structure is best represented by the two factor model – narcissistic pride and humble pride.
  In line with earlier research, the CFA supported that narcissistic pride and humble pride were two separate but related factors. The factor loadings indicated that the factors were well defined by their corresponding items, and all remaining loadings were significant, positively linked to the corresponded latent factor, and substantial in size ranging from 0.45 to 0.74.
  Therefore, a four item Narcissistic Pride scale comprised the average of ratings to the following items: (a) feel superior over your co-workers; (b) feel like putting co-workers down; (c) feel like you had authority over your co-workers; and (d) feel dominant over your co-workers (Mean = 1.36; SD = 0.45; Cronbach’s alpha = 0.68).
  The Humble Pride scale comprised the average of ratings to eight items: (a) willing to take personal responsibility for any mistakes you made along the way; (b) be modest with your co-workers; (c) feel proud of yourself without being arrogant; (d) be considerate to your co-workers’ comments on this solution; (e) respect all co-workers irrespective of status; (f) feel a sense of achievement without being arrogant; (g) feel good about other co-workers who helped solve the problem; and (h) respect the contribution of other co-workers to solving the problem (Mean = 3.19; SD = 0.48; Cronbach’s alpha = 0.85).

  A.1.3 Dependent variables: Bullying and victimisation

  The measure of workplace bullying is a modified version of that developed by Quine (1999). Respondents were presented with 17 ways of bullying others. They used a four-point rating scale (1=“never,” 2=“a few times,” 3=“sometimes,” and 4=“often”) to indicate first the frequency with which they had treated others this way in the past year, and next, the frequency with which they had been treated this way by others in the past year (see below for a complete list of items).
  Based on modification indices, most error terms were co-varied. In addition, two items from each scale⁴x‘Physical threats’ and ‘Threats to property’ were removed from the Victimisation scale, and ‘Shifting of goal posts without telling’ and ‘Removal of areas of responsibility without consultation’ were removed from the Bullying scale. were removed due to their very low factor loadings.
  Findings of the CFAs indicated a good fit to the data of the one-factor model for bullying (Model fit: χ² = 42.05; df = 32; p < 0.11; χ²/df = 1.31; AGFI = 0.981; CFI = 0.999; TLI = 0.995; RMSEA = 0.017; SRMR = 0.003). The structural model fit indices also indicated good model fit for victimisation (Model fit: χ² = 42.84; df = 38; p < 0.27; χ²/df = 1.12; AGFI = 0.984; CFI = 0.999; TLI = 0.998; RMSEA = 0.011; SRMR = 0.006).
  Based on the CFA analyses, the following two scales were formed:
  

  1. Workplace Bullying scale scores were formed by averaging responses to bullying others over the fifteen items (Mean = 1.19; SD = 0.29; Cronbach’s alpha = 0.88).

  2. Victimisation scale scores were formed by averaging responses to being bullied over the fifteen items (Mean = 1.46; SD = 0.53; Cronbach’s alpha = 0.93).

  Workplace bullying and victimisation items

  Statements about what could happen in your workplace. In the past year, how often has this happened?
  1	Making inappropriate jokes (B)*(V)*
  2	Teasing (B)(V)
  3	Freezing out/ignoring/ excluding (B)(V)
  4	Destructive innuendo and sarcasm (B)(V)
  5	Unjustified criticism of work (B)(V)
  6	Attempts to humiliate in front of co-workers (B)(V)
  7	Unreasonable pressure to produce work (B)(V)
  8	Shifting of goal posts without telling (V)
  9	Constant undervaluing of efforts (B)(V)
  10	Attempts to belittle work (B)(V)
  11	Intimidatory use of discipline procedures (B)(V)
  12	Verbal threats (B)(V)
  13	Removal of areas of responsibility without consultation (V)
  14	Attempts to demoralise as a person (B)(V)
  15	Unreasonable refusal of applications for leave, training, promotion (B)(V)
  16	Physical threats (B)
  17	Threat to property (B)

  * B means this item is part of the Bullying scale; V means this item is part of the Victimisation scale.

Appendix 2

Zero-order correlations and descriptive statistics for all predictor and dependent variables (n = 1,045)

No	Variables	1	2	3	4	5	6	7	8	9	10	11	12
1	Gender	–
2	Age	0.10***	–
3	Education	–0.04(ns)	–0.10***	–
4	Personal Income	0.27***	0.16***	27***	–
5	Forgive offender	–0.11***	0.06 (ns)	0.06 (ns)	–0.04 (ns)	–
6	See offender rehabilitate	–0.11***	0.01 (ns)	–0.04	–0.09**	0.46***	–
7	Distrust	0.20***	0.05 (ns)	–0.20***	–0.03 (ns)	–0.18***	–0.13***	–
8	Shame acknowledgment	–0.08**	–0.08**	0.07*	0.00 (ns)	0.12***	0.09**	–0.02 (ns)	–
9	Shame displacement	–0.01 (ns)	–0.07*	0.07*	–0.01 (ns)	0.03 (ns)	0.01 (ns)	0.06*	0.40***	–
10	Humble pride	–0.09**	0.11***	0.03 (ns)	0.02 (ns)	0.17***	0.12***	–0.16***	0.16***	–0.13***	–
11	Narcissistic pride	0.12***	–0.13***	0.08**	0.12***	–0.12***	–0.07*	0.16***	0.19***	0.32***	–0.15***	–
12	Bullying	0.18***	–0.04 (ns)	–0.03 (ns)	0.06*	–0.12***	–0.11***	0.21***	0.10***	0.23***	–0.14***	0.37***	–
13	Victimisation	0.04 (ns)	–0.10**	–0.02 (ns)	0.03 (ns)	–0.07*	–0.02 (ns)	0.20***	0.20***	0.26***	–0.03 (ns)	0.24***	0.50***

Descriptive Statistics	1	2	3	4	5	6	7	8	9	10	11	12	13
Mean	na	41.7	na	39,842	3.36	3.44	2.60	2.04	1.37	3.20	1.37	1.19	1.46
Standard Deviation	na	12.3	na	31,044	1.04	0.98	0.60	0.64	0.47	0.47	0.45	0.30	0.53
Alpha Reliability	na	na	na	na	na	na	0.67	0.85	0.71	0.85	0.68	0.88	0.93

***p < 0.001, **p < 0.01, *p < 0.05

Noten

• * Contact author: Valerie.Braithwaite@anu.edu.au.

• 1 See, for example, Australian Public Service Commission. (2011). Respect: promoting a culture free from harassment and bullying in the APS. 4th ed. Canberra, ACT: Australian Public Service Commission.

• 2 We used the following indices to assess the fit of the model to the data: the chi-square (χ²) statistic and the related degrees of freedom (df), ratio of χ² and df, the adjusted goodness of fit index (AGFI), the comparative fit index (CFI), the Tucker-Lewis index (TLI), the root mean square error of approximation (RMSEA), and the standardised root mean square residual (SRMR). In general, a χ²/df value <3.00 and AGFI, CFI and TLI values >.90 indicate an acceptable fit (with values >.95 being ideal (Brown, 2006). Further, RMSEA and SRMR values ≤.08 indicate a reasonable fit to the data, whereas values ≤.05 indicate excellent fit (Hu & Bentler, 1999). Note that the χ² statistic is very sensitive to sample size and hence is no longer relied upon as a basis for acceptance or rejection of a model (Schermelleh-Engel, Moosbrugger & Müller, 2003; Vandenberg, 2006). As a result, the use of multiple fit indexes has developed to provide a more holistic view of goodness of fit, taking account not only of sample size but also of model complexity and other relevant issues of the study such as multivariate normality.

• 3 ‘Pretend that nothing had happened’ representing shame displacement.

• 4 ‘Physical threats’ and ‘Threats to property’ were removed from the Victimisation scale, and ‘Shifting of goal posts without telling’ and ‘Removal of areas of responsibility without consultation’ were removed from the Bullying scale.