ویژگیهای شخصیتی زورگویان محل کار اغلب توسط قربانیان آنها به اشتراک گذاشته می شود: نیمه تاریکی برای قربانیان وجود دارد؟

Abstract In this exploratory study, we examined the extent to which both workplace bullies and victims possess bully-typifying traits, using a 22-item scale that simultaneously measures perpetrators and targets of negative workplace acts. Participants were 224 Canadian university students aged 18–47 with prior work experience. Bivariate correlational analyses determined that bullying others was positively associated with measures of Machiavellianism, narcissism, psychoticism, aggression, and disinhibition. Being a victim was positively associated with the same Machiavellianism, narcissism, psychoticism, and aggression measures. Hierarchical regression analyses indicated that an “alternative dark triad” of Machiavellianism, narcissism, and psychoticism related significantly to bullying scores; while psychoticism and Machiavellianism related significantly to victim scores. Aggression and sensation seeking measures failed to account for significant variance in bully or victim scores beyond the triad variables. The vast majority of bullies (89.7%) and many victims (41.7%) were bully/victims, operationally defined as being both perpetrators and targets at least once per week in the last 6 months. Researchers and employers would do well to recognize the presence of bully/victims in their efforts to understand and reduce workplace bullying

1. Introduction An understanding of workplace bullying may help reduce the health and emotional costs for victims. Bullied employees are at increased risk of physical and psychological problems including depression, psychosomatic symptoms, post-traumatic stress, and coronary heart disease (Einarsen et al., 2009, Lee and Brotheridge, 2006 and Lutgen-Sandvik et al., 2007). Workplace bullying has been cited with prevalence rates of 46.8%, 24.1%, and 15.8%, at least once weekly in American, Finnish, and Danish studies, respectively (Lutgen-Sandvik et al., 2007). The Negative Acts Questionnaire-Revised (NAQ-R) is a widely-used measure of bullying victimization which asks employees to report the extent to which they have experienced a variety of negative acts at work. We modified the victim-oriented NAQ-R into a Perpetrator-Target Scale (PTS), creating “bully” and “victim” subscales, which simultaneously measure both perpetrators and targets of bullying. Bullies and victims appear to share personality traits that are more typically attributed to bullies. Scattered observations have found that victims possess such bully-typifying traits as disagreeableness, dominance, and aggression (Archer et al., 2007, Glaso et al., 2009 and Olweus, 1993). Bullies and victims may share personality traits because many are both perpetrators and targets of negative acts. In fact, Lee and Brotheridge (2006) found that 83% of bullies were also victims, and 50% of victims were also bullies, on a once weekly basis. Because victims are often bullies, we suspect that they possess some of the stable and enduring personality traits consistent with the bully role. In addition, bullying researchers have recognized that the personalities of some persons are sufficiently provocative that others may respond by bullying them (Olweus, 1993 and Matthiesen and Einarsen, 2007). Andersson and Pearson (1999), for example, have described the process whereby employees feel justified in responding aggressively to annoying co-workers whose behavior and attitudes fall outside of social norms. Many provocative personalities are also bully/victims, as they may be perpetrators as well as targets (Matthiesen & Einarsen, 2007). The provocative traits of victims range from the aggressive, to those deemed “annoying” or socially inappropriate (Andersson and Pearson, 1999 and Matthiesen and Einarsen, 2007), labels that are likewise applicable to the traits of bullies. As an example, the bully-typifying trait verbal aggression may be directly associated with the bullying of others (Parkins, Fishbein, & Ritchey, 2006). It is also, however, associated with argumentativeness (Diamond, 2005), a possible “annoying” factor that may provoke others to bully those who score highly on the trait. In this study, we have compiled a number of personality traits that have been either empirically or theoretically linked with bullying others; we expect victims to share these traits. Machiavellianism, narcissism, and subclinical psychopathy, the so-called “dark triad” of personality, have been associated with bullying. Psychopathy was most strongly associated with bullying, followed by Machiavellianism and narcissism in a recent workplace bullying study (Baughman, Dearing, Giammarco, & Vernon, 2012). Those high in Machiavellianism are believed to “employ aggressive, exploiting, and devious behavior to achieve personal and organizational goals,” with little thought for the welfare of others (Ricks & Fraedrich, 1999, p. 197–198). High Machiavellian employees have self-reported as more comfortable with violating others’ rights than low scorers (Winter, Stylianou, & Giacalone, 2004). Narcissistic individuals are grandiose, with a sense of entitlement and superiority which may cause them to feel entitled to treat others badly. Narcissism scores in undergraduate students have been found to be positively associated with self- and observer-reports of antisocial behavior, on a measure which included bullying items (Paulhus, Robins, Trzesniewski, & Tracy, 2004). Those high in subclinical psychopathy are more likely to perform antisocial behaviors and are typically cold, manipulative, impulsive, thrill-seeking, and distrustful (LeBreton, Binning, & Adorno, 2006). Eysenckian psychoticism was derived as a complex supertrait in the “big three” model that addresses the major factors of personality (Eysenck & Eysenck, 1994). This measure has been characterized as a construct related to, or on continuum with, subclinical psychopathy (Corr, 2010 and Zuckerman, 1991). Considerable theory and research supports the instrument’s utility for normal populations, involving hundreds of studies in many countries. This broadly heterogenous scale addresses a large range of subtraits, many of which are of relevance for bullying research. The high scorer is unempathetic, manipulative, irresponsible, risk-taking (disregarding danger), inhumane, impulsive, insensitive, hostile (even to own kin), aggressive (even to loved ones), unconventional (with a liking for odd and unusual things), under socialized, disrespectful of laws, manners, and traditions, likes to upset and make fools of others, is troublesome, and expresses antisocial attitudes and behaviors (Eysenck & Eysenck, 1994). A junior version of this scale has correlated positively with bullying in school children (Connolly & O’Moore, 2003) Aggression and sensation seeking are also of interest to bullying research. Workplace bullying has correlated significantly with Buss-Perry Aggression subscales in employed college students (Parkins et al., 2006). Zuckerman (2007) concluded that sensation seeking was associated with aggression and antisocial behavior, finding the disinhibition and boredom susceptibility components most related to aggression scale scores. Bullies, like young offenders (Gudjonsson & Sigurdsson, 2007), may perpetrate for excitement, to relieve monotony, or through low impulse control. In this study, we examine the extent to which victims and bullies possess bully typifying traits. It is hypothesized that the bully-typifying traits Machiavellianism, narcissism, Eysenckian psychoticism, aggression, and sensation seeking will associate positively with both bully and victim scores.

3. Results 3.1. Demographics There were 158 females (70.5%), and 66 males (29.5%) included in the study. The mean age was 21.45 years (SD = 4.89), with a range of 18–47 years. Participants self-reported working an average of 29.83 hours weekly, had been with their present employer an average of 2.24 years, and had been in their current occupation an average of 2.34 years. Some 40.2% of the participants reported being employed full time. Demographic variables such as hours worked per week, years with current employer and occupation, occupation type age, employment status (i.e., full versus part-time), educational level, ethnicity, recruitment group (in terms of university and discipline/degree program), body height, and marital status were not significantly related to scores on the bully or victim subscales. Body weight was significantly correlated with the bully subscale, however, and remained so after controlling for gender (r = .14, p = .05). Separate bivariate analyses by gender showed that this association characterized the female (r = .19, p = .02), but not the male, subgroup. 3.2. Perpetrator-Target Scale (PTS) Roughly 37.5% of participants who completed the PTS reported being victimized by negative acts at least once per week in the past 6 months. Approximately 17.4% reported perpetrating negative acts upon others by the same criteria. The mean scores on the bully and victim subscale items ranged from 1.04 to 1.70, and from 1.10 to 2.07, respectively. The overall subscale means and standard deviations are reproduced in Table 2. Table 2. Means, standard deviations (SD), and Pearson product-moment correlations of study variables. Variable Mean (SD) 1 2 3 4 5 6 7 8 9 10 1. Victim subscale 31.62 (9.50) 2. Bully subscale 28.09 (6.10) 0.62⁎⁎⁎ 3. Machiavellianism 2.02 (2.20) 0.26⁎⁎⁎ 0.44⁎⁎⁎ 4. Narcissism 2.77 (1.98) 0.19⁎⁎ 0.38⁎⁎⁎ 0.53⁎⁎⁎ 5. Psychoticism 4.28 (2.75) 0.32⁎⁎⁎ 0.35⁎⁎⁎ 0.39⁎⁎⁎ 0.33⁎⁎⁎ 6. Anger 2.08 (1.04) 0.19⁎⁎ 0.29⁎⁎⁎ 0.25⁎⁎⁎ 0.11 0.33⁎⁎⁎ 7. Hostility 2.55 (0.98) 0.21⁎⁎⁎ 0.22⁎⁎⁎ 0.24⁎⁎⁎ −0.05 0.24⁎⁎⁎ 0.58⁎⁎⁎ 8. Verbal aggression 2.55 (0.92) 0.25⁎⁎⁎ 0.26⁎⁎⁎ 0.31⁎⁎⁎ 0.32⁎⁎⁎ 0.32⁎⁎⁎ 0.45⁎⁎⁎ 0.29⁎⁎⁎ 9. BSSS Sensation seeking total 26.50 (5.78) 0.07 0.08 0.23⁎⁎⁎ 0.17⁎ 0.44⁎⁎⁎ 0.25⁎⁎⁎ 0.14⁎ 0.09 10. Boredom susceptibility 3.46 (0.85) 0.06 0.10 0.19⁎⁎ 0.14⁎ 0.27⁎⁎⁎ 0.21⁎⁎⁎ 0.09 0.15⁎ 0.63⁎⁎⁎ 11. Disinhibition 2.96 (1.02) 0.12 0.19⁎⁎ 0.27⁎⁎⁎ 0.23⁎⁎⁎ 0.45⁎⁎⁎ 0.30⁎⁎⁎ 0.22⁎⁎⁎ 0.19⁎⁎ 0.78⁎⁎⁎ 0.36⁎⁎⁎ ⁎ p < .05. ⁎⁎ p < .01. ⁎⁎⁎ p < .001. Table options Significantly higher scores were obtained on the victim (M = 31.62, SD = 9.50) than bully subscales (M = 28.09, SD = 6.10), t (2 2 3) = 7.07, p < .001, d = .47. There were no significant differences between males and females in terms of either bully, t (2 2 2) = 1.21, p = .23, or victim scores, t (2 2 2) = 0.64, p = .52. A substantive correlation, r = .62 (p < .001), was obtained between the bully and victim subscales, suggesting that bullying others and being a victim were strongly associated (see Table 2). Both the bully and victim subscales of the PTS possess high internal consistency (α = .84 and .89, respectively), with alpha failing to improve upon removal of any of the 22 items. The construct validity of the bully subscale was supported through associations with general aggression measures (e.g., BPAQ-SF Anger, Hostility, and Verbal Aggression), and personality traits previously observed as positive bullying correlates (Machiavellianism, narcissism). In addition, we have obtained a correlation of r = .62 (p < .001) between the PTS bully subscale and the 6-item Workplace Bullying Scale ( Parkins et al., 2006), another measure of bullying others (Linton & Power, unpublished results). The notion of a positive relationship between bully and victim scores was supported by a significant correlation between the PTS victim subscale and the Workplace Bullying Scale (r = .34, p < .001). 3.3. Personality variables The mean scores, standard deviations, and Pearson product-moment correlations are presented for the PTS and personality variables in Table 2. Consistent with study hypotheses, Machiavellianism, narcissism, EPQ-R Psychoticism, the BPAQ-SF aggression subscales, and BSSS Disinhibition correlated significantly with the PTS bully subscale. Hypotheses for the overall Brief Sensation Seeking Scale (BSSS) and Boredom Susceptibility were not supported for bullies. Also consistent with study hypotheses, Machiavellianism, narcissism, EPQ-R Psychoticism, and the BPAQ-SF aggression subscales correlated significantly with the PTS victim subscale. Hypotheses concerning BSSS overall scale scores, Disinhibition and Boredom Susceptibility were not supported for victims. Correlations with the bully-typifying traits tended to be of lower magnitude for victims than bullies. Psychoticism, hostility, and verbal aggression, however, were notable as their correlations with the victim and bully subscales were very similar (psychoticism obtained 0.32 and 0.35, hostility obtained 0.21 and 0.22, and verbal aggression obtained 0.25 and 0.26, for victims and bullies, respectively). Correlations between the bully-typifying traits, and PTS bully subscales created to correspond with NAQ-R-derived factors (work-related bullying, personal bullying and physically intimidating bullying), were similar to those observed between the bully-typifying traits and the bully subscale of the PTS. For each subscale, correlations with Machiavellianism, narcissism and psychoticism exceeded those for the BPAQ aggression variables, which in turn exceeded those for the BSSS sensation seeking variables. 3.4. Results of hierarchical regression on the bully subscale To examine the contributions of other subtypes of bully-typifying variables (BPAQ-SF aggression, BSSS sensation seeking), after Machiavellianism, narcissism, and psychoticism had been entered, hierarchical multiple regression analyses were utilized. Separate analyses were performed for the bully and the victim subscales. Tolerance values of 0.77–0.92 (>0.10) and variance inflation factors of 1.09–1.29 (<10.00) for the regressed variables suggest that error due to multicolinearity is unlikely. Step 1 of a 3-step hierarchical regression procedure predicted bullying scores from Machiavellianism, narcissism, and psychoticism. R2 for step 1 was .29, a value that was highly significant, F(3, 217) = 28.93, p < .001. In step 2, the BPAQ-SF aggression variables, and in step 3, the sensation seeking variables (boredom susceptibility and disinhibition) failed to significantly increase the proportion of explained variance in the bullying variable, with R2 changes of .02 and .003, respectively. Table 3 shows regression coefficients, along with significance tests. Table 3. Hierarchical regression analyses predicting bully and victim subscale scores.. Bully subscale scores Victim subscale scores β β Predictor Step 1 Machiavellianism 0.32⁎⁎⁎ 0.16⁎ Narcissism 0.15⁎ 0.04 Psychoticism 0.20⁎⁎ 0.24⁎⁎⁎ Step 2 Machiavellianism 0.28⁎⁎⁎ 0.11 Narcissism 0.18⁎ 0.06 Psychoticism 0.16⁎ 0.20⁎⁎ Anger 0.1 0 Hostility 0.06 0.13 Verbal aggression 0.01 0.1 Step 3 Machiavellianism 0.28⁎⁎⁎ 0.11 Narcissism 0.17⁎ 0.07 Psychoticism 0.18⁎ 0.23⁎⁎ Anger 0.1 0.13 Hostility 0.07 0.13 Verbal aggression −0.01 0.1 Boredom susceptibility 0.05 0 Disinhibition −0.06 −0.09 Note: Bully subscale: R2 = .29 ⁎⁎⁎ for Step 1; ΔR2 = .02 for Step 2; ΔR2 = .003 for Step 3. Victim subscale: R2 = .13 ⁎⁎⁎ for Step 1; ΔR2 = .03 for Step 2; ΔR2 = .006 for Step 3. ⁎ p < .05. ⁎⁎ p < .01. ⁎⁎⁎ p < .001. Table options 3.5. Results of hierarchical regression on the victim subscale Step 1 of a 3-step hierarchical regression procedure predicted victim scores from Machiavellianism, narcissism, and psychoticism. R2 for step 1 was .13, a value that was highly significant, F(3, 217) = 11.10, p < .001. Machiavellianism and psychoticism were significant predictors in this model. In step 2, the BPAQ-SF aggression variables, and in step 3, the sensation seeking variables (boredom susceptibility and disinhibition) failed to significantly increase the proportion of explained variance in the bullying variable, with R2 changes of .03 and .006, respectively. See Table 3. 3.6. Bully/victim analysis To investigate a possible explanation for the expression of bully-typifying traits in victims, the prevalence of bully/victims was calculated. Bully/victims were defined as those who had been both perpetrators and targets of negative events at least once weekly during the past 6 months. Bully/victims comprised 15.6% of the overall study sample. Here, we found that a substantive 41.7% of victims and 89.7% of bullies were also bully/victims. We then examined whether bully/victims possess higher mean scores than non-bully/victims on the bully-typifying personality traits Machiavellianism, narcissism, and EPQ-R Psychoticism (selected because of their predictive value in the correlational and regression analyses). Bully/victims were categorized as those who scored above the median on both the PTS bully and victim subscales, while non-bully/victims were categorized as those who scored below the median on at least one of the two subscales. Means and standard deviations for bully/victims and non-bully/victims were calculated at an adjusted alpha level of .016 (.05/3) to minimize type 1 error. Table 4 demonstrates significantly higher mean scores for bully/victims on each trait. Table 4. Means and standard deviations for Machiavellianism, narcissism, and psychoticism by bully/victim status. Bully/victim M (SD) Non-bully/victim M (SD) t Cohen’s d Machiavellianism 2.95 (2.64) 1.49 (1.71) 4.59⁎⁎⁎ 0.63 Narcissism 3.32 (2.19) 2.43 (1.78) 3.19⁎⁎ 0.44 Psychoticism 5.20 (2.72) 3.71 (2.63) 4.06⁎⁎⁎ 0.56 ⁎⁎ p < .01. ⁎⁎⁎ p < .001.