British Journal of Management, Vol. 21, 438452 (2010) DOI: 10.1111/j.1467-8551.2009.00672.

The Prevalence of Destructive Leadership Behaviour

Merethe Schanke Aasland, Anders Skogstad, Guy Notelaers, Morten le Einarsen Birkeland Nielsen and Sta
Department of Psychosocial Science, Christies Gate 12, University of Bergen, N-5015 Bergen, Norway Corresponding author email: merethe.aasland@psysp.uib.no
This study investigates the prevalence of the four types of destructive leadership behaviour in the destructive and constructive leadership behaviour model, in a representative sample of the Norwegian workforce. The study employs two estimation methods: the operational classication method (OCM) and latent class cluster (LCC) analysis. The total prevalence of destructive leadership behaviour varied from 33.5% (OCM) to 61% (LCC), indicating that destructive leadership is not an anomaly. Destructive leadership comes in many shapes and forms, with passive forms prevailing over more active ones. The results showed that laissez-faire leadership behaviour was the most prevalent destructive leadership behaviour, followed by supportivedisloyal leadership and derailed leadership, while tyrannical leadership behaviour was the least prevalent destructive leadership behaviour. Furthermore, many leaders display constructive as well as destructive behaviours, indicating that leadership is not either constructive or destructive. The study contributes to a broader theoretical perspective on what must be seen as typical behaviour among leaders.

Introduction
So far, research on destructive forms of leadership has mainly focused on its detrimental eects on subordinates (Bies and Tripp, 1998; Tepper, 2000; Zellars, Tepper and Duy, 2002). However, few studies have investigated the prevalence of such destructive leadership in contemporary working life (e.g. Schat, Frone and Kelloway,
The present project is a collaborative project between the University of Bergen and Statistics Norway, which collected the data. The project was made possible by joint grants from two Norwegian employer associations (the Confederation of Norwegian Enterprise and the Norwegian Association of Local and Regional Authorities), and the Norwegian government (the National Insurance Administration) and their FARVE programme. We would like to thank Bengt Oscar Lagerstrm and Maria Hstmark of Statistics Norway and Stig Berge Matthiesen of the Faculty of Psychology, University of Bergen, for their contribution to the data collection.

2006). While some researchers claim that destructive or abusive leadership constitutes a low baserate phenomenon (e.g. Aryee et al., 2007), others believe it to be a substantial problem in many organizations, in terms of both its prevalence and consequences (Burke, 2006; Hogan, Raskin and Fazzini, 1990). In organizational climate research from the mid-1950s to 1990, 60%75% of all employees typically reported that the worst aspect of their job was their immediate supervisor (Hogan, Raskin and Fazzini, 1990). In the USA, job pressure has been cited in 75% of workers compensation claims in which mental stressors were the main cause of absenteeism, and 94% of those claims were allegedly caused by abusive treatment by managers (Wilson, 1991). Thus, a growing body of evidence documents that leaders behave in a destructive manner, be it towards their subordinates (Bies and Tripp, 1998; Tepper, 2000, 2007), towards the organization itself or towards both (Kellerman, 2004; Vredenburgh

r 2009 British Academy of Management. Published by Blackwell Publishing Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main Street, Malden, MA, 02148, USA.

Destructive Leadership Behaviour and Brender, 1998). The prevalence of destructive leadership has serious implications for theories on leadership and for measures of leadership, as well as for the recruitment, training and development of leaders. The aim of the present study is thus to contribute to the growing body of literature on destructive leadership by investigating the prevalence of four forms of destructive leadership behaviour derived from the Destructive and Constructive Leadership (DCL) behaviour model (Einarsen, Aasland and Skogstad, 2007) employing a representative sample of the Norwegian workforce. Conceptualizations of destructive leadership Many concepts have been used to describe destructive forms of leadership, such as abusive supervision (Tepper, 2000) and petty tyranny (Ashforth, 1994), referring to leaders who behave in a destructive manner towards subordinates, by intimidating subordinates, belittling or humiliating them in public or exposing them to nonverbal aggression (Aryee et al., 2007). Concepts such as authoritarian (Adorno et al., 1950; Bass, 1990a), Machiavellian (Christie and Geis, 1970), autocratic (Kipnis et al., 1981), narcissistic leadership (Kets de Vries and Miller, 1985) and personalized charismatic leadership (House and Howell, 1992) emphasize similar but not overlapping behaviours. However, these concepts mainly focus on control and obedience, and less on the abusive aspect of leadership. Leaders may also behave destructively in a way that primarily aects the organization (Kellerman, 2004; Lipman-Blumen, 2005), potentially leading to negative consequences for the execution of tasks, quality of work, eciency and relations with customers and clients (Padilla, Hogan and Kaiser, 2007). Concepts frequently used to describe such behaviour are awed leadership (Hogan, 1994), derailed leadership (McCall and Lombardo, 1983; Shackleton, 1995), the dark side of leadership (Conger, 1990), toxic leadership (Lipman-Blumen, 2005) and impaired managers (Lubit, 2004). Such leaders neglect, or even actively prevent, goal attainment in the organization by, for example, sabotaging subordinates task execution, by working towards alternative goals than those of the organization (Conger, 1990), by stealing resources such as materials, money or time, or by encouraging
r 2009 British Academy of Management.

439 employees to engage in such activities (Altheide et al., 1978). Considering the breadth of the concepts used to describe destructive leaders, it seems clear that destructive leadership is not one type of leadership behaviour, but instead involves a variety of behaviour. Taking this diversity into account, the present study uses the overarching concept of destructive leadership, dened as systematic and repeated behaviour by a leader, supervisor or manager that violates the legitimate interest of the organization by undermining and/or sabotaging the organizations goals, tasks, resources, and eectiveness and/or the motivation, wellbeing or job satisfaction of subordinates (Einarsen, Aasland and Skogstad, 2007, p. 208). Hence, destructive leadership is about systematically acting against the legitimate interest of the organization, whether by abusing subordinates or by working against the attainment of the organizations goals, including any illegal behaviour. The denition emphasizes repeated destructive behaviour as opposed to a single act such as an isolated outburst of anger or spontaneous misbehaviour. However, if mistakes or outbursts of anger become repeated, they represent destructive leadership according to the denition irrespective of their intentions or antecedents. Furthermore, destructive leadership is about behaviour that violates, or is in opposition to, what is considered to be the legitimate interest of the organization. Including legitimate interest is in accordance with Sackett and DeVores (2001) denition of counterproductive workplace behavior, narrowing what an organization may expect from its leaders to what must be seen as legitimate, legal, reasonable and justiable behaviour in a given cultural setting. Hence, what is perceived as destructive behaviour may vary between cultures and societies and also over time (Einarsen, Aasland and Skogstad, 2007). In accordance with the above denition, the DCL model (see Figure 1) describes four main kinds of destructive leadership behaviour targeting either subordinates and/or the organization (Einarsen, Aasland and Skogstad, 2007). Categorizing leadership behaviour as subordinateoriented and organization-oriented is not new. However, while many models of leadership behaviour, such as the managerial grid developed by Blake and Mouton (1985) and the full range of leadership model (Avolio, 1999; Bass and Riggio,

440
Pro-subordinate Supportivedisloyal Leadership Behaviour Anti-organization Derailed Leadership Behaviour Laissez-faire Leadership Behaviour Tyrannical Leadership Behaviour Constructive Leadership Behaviour

M. S. Aasland et al.

Pro-organization

Anti-subordinate
Figure 1. A model of destructive leadership behaviour

2006), are based on the assumption that leadership behaviour can be seen on a continuum from low to high with regard to constructive leadership behaviour, the present model views leadership behaviour on a continuum from highly anti to highly pro (Aasland, Skogstad and Einarsen, 2008). Hence, the subordinate dimension describes leadership behaviour ranging from anti-subordinate behaviour to pro-subordinate behaviour. Anti-subordinate behaviour illegitimately undermines or sabotages the motivation, well-being or job satisfaction of subordinates, involving behaviour such as harassment and mistreatment of subordinates (Einarsen, Aasland and Skogstad, 2007). Pro-subordinate behaviour fosters the motivation, well-being and job satisfaction of subordinates, including taking care of and supporting them in accordance with organizational policies. Organization-oriented behaviour may also range from anti-organization behaviour to pro-organization behaviour, where the former violates the legitimate interest of the organization by working in opposition to the organizations goals, values and optimal use of resources, by stealing from the organization, by sabotaging the organizations goals, or even by being involved in corruption (Einarsen, Aasland and Skogstad, 2007). Anti-organizational behaviour can also be described as counterproductive workplace behaviour directed at the organization (Fox and Spector, 1999; Sackett and DeVore, 2001). Proorganizational behaviour is about working to-

wards the fullment of the organizations goals, setting clear and unambiguous objectives, making or supporting strategic decisions and implementing legitimate organizational change. By crosscutting the two dimensions, the DCL model presents ve categories of leadership behaviour, one of which is constructive, three of which are actively destructive tyrannical, derailed, and supportivedisloyal leadership behaviour (Einarsen Aasland and Skogstad, 2007) and one of which is passive: laissez-faire leadership, situated in the middle of the proposed model. Constructive leadership is in accordance with the legitimate interest of the organization, showing both pro-subordinate and pro-organization behaviour to some degree. Constructive leadership is about displaying behaviour that involves supporting and enhancing the goal attainment of the organization, making optimal use of organizational resources, as well as enhancing the motivation, well-being and job satisfaction of subordinates. Such leadership behaviour can also be described using concepts such as transactional (Bass et al., 2003), transformational (Bass et al., 2003), charismatic (e.g. Conger and Kanungo, 1987) and empowering leadership (e.g. Conger and Kanungo, 1988). However, the focus of the present paper is on destructive leadership behaviours. Tyrannical leadership behaviour is about displaying pro-organizational behaviour combined with anti-subordinate behaviour. Strictly speaking, such leaders may behave in accordance with
r 2009 British Academy of Management.

Destructive Leadership Behaviour the legitimate goals, tasks and strategies of the organization. However, they typically obtain results not through, but at the expense of, subordinates (Ashforth, 1994; Ma, Karri and Chittipeddi, 2004). Tyrannical leaders may humiliate, belittle and manipulate their subordinates in order to get the job done. Because tyrannical leaders may behave constructively in terms of organization-oriented behaviour while at the same time displaying anti-subordinate behaviour, subordinates and superiors may evaluate the leaders behaviour quite dierently. What upper management may see as a strong focus on task completion may at the same time be seen by subordinates as abusive leadership or even bullying (Einarsen, Aasland and Skogstad, 2007). Derailed leadership is about displaying both anti-organizational and anti-subordinate behaviour. Such leaders may bully, humiliate, manipulate or deceive, while simultaneously engaging in anti-organizational behaviour such as absenteeism, fraud or otherwise stealing resources from the organization (Aasland, Skogstad and Einarsen, 2008; McCall and Lombardo, 1983). Conger (1990) focuses on similar themes in his study of the dark side of leadership, in which he recognizes that leaders may use their charismatic qualities for personal gain and abusively turn against what is good for both followers and the organization. Supportivedisloyal leadership consists of prosubordinate behaviour combined with anti-organizational behaviour. Such leaders motivate and support their subordinates, while simultaneously stealing resources from the organization, be it materials, time or nancial resources (Altheide et al., 1978; Ditton, 1977). Supportivedisloyal leaders may give employees more benets then they are entitled to at the expense of the organization, encourage low work ethics and misconduct and lead their subordinates to be inecient, or towards other goals than those of the organization, all of this while behaving in a comradely and supportive manner. They may also commit embezzlement or fraud, or encourage subordinates to enrich themselves through such anti-organizational behaviour (Einarsen, Aasland and Skogstad, 2007). This form of leadership behaviour has some features in common with the leadership style that Blake and Mouton (1985) termed country club management, as both forms of leadership behaviour
r 2009 British Academy of Management.

441 reect an overriding concern with establishing camaraderie with subordinates and ensuring their well-being. However, while country club management involves a minimum of focus on production and eciency, thus being low on the organization-oriented dimension, leaders who behave in a supportive but disloyal manner behave destructively towards the organization, thus behaving in an anti-organizational manner (Aasland, Skogstad and Einarsen, 2008). Hence, the absence of constructive leadership behaviour is not the same as the presence of destructive leadership behaviour (Kelloway, Mullen and Francis, 2006). However, destructive leadership is not necessarily limited to such active and manifest behaviour as described above. Buss (1961) describes aggressive behaviour along three principal axes, namely physical versus verbal, active versus passive and direct versus indirect aggression. Consequently, destructive leadership behaviour may also include passive and indirect behaviour (Skogstad et al., 2007). Kelloway and colleagues (2005) also acknowledge this in their study of poor leadership, in which they dierentiate between an active, abusive leadership style and a passive one. Avoidant or passive leadership, which is also referred to as laissez-faire leadership (Bass, 1990b), represents a leadership style in which the leader has been appointed to and still physically occupies the leadership position, but in practice has abdicated the responsibilities and duties assigned to him or her (Lewin, Lippitt and White, 1939). Such leaders may avoid decisionmaking, show little concern for goal attainment and seldom involve themselves with their subordinates, even when this is necessary (Bass, 1990b). Ashforth (1994) emphasizes the importance of passive destructive behaviour in his conceptualization of the petty tyrant, including lack of consideration and discouraging initiative as two of six dimensions. Thus, the systematic absence of positive behaviour is conceptualized as destructive leadership behaviour. Independently of the causes of passive or laissez-faire leadership behaviour, be it a result of incompetence, lack of knowledge or strategic intent to harm, it clearly violates the legitimate interest of the organization as well as legitimate expectations of subordinates, and it may thus harm both the organization and the subordinates (Frischer and Larsson, 2000; Hinkin and Schriesheim, 2008; Skogstad et al., 2007).

442 Aim of the study A small but growing body of studies exists on destructive forms of leadership behaviour. However, these studies are mostly limited to the characteristics of such destructive leadership and its eects on subordinates. Apart from two studies investigating the prevalence of leadership aggression (Hubert and van Veldhoven, 2001; Schat, Frone and Kelloway, 2006), we know little about how prevalent various forms of destructive leadership behaviour are. Such knowledge is of great importance, especially since eorts to develop eective interventions against such behaviour may depend on the prevalence of the phenomenon (Zapf et al., 2003). Moreover, further theoretical developments specically relating to destructive leadership, as well as to leadership in general, depend on an estimate of the prevalence of destructive leadership behaviour. Nuanced information in this regard may alter our perception of leadership as a phenomenon and lay the foundation for how much attention should be devoted to this aspect of leadership in future leadership training and development (Burke, 2006). Hence, the aim of the present study is to investigate, on the basis of a representative sample of subordinates, the prevalence of the four forms of destructive leadership behaviour laid out in the DCL model. Measures

M. S. Aasland et al.

Method
Procedure/Sample Questionnaires were sent by regular mail to a representative sample of 4500 employees, randomly drawn from the Norwegian Central Employee Register during spring 2005, with two reminders. The sampling criteria were employees between 18 and 65 years of age, employed during the last six months in a Norwegian company with ve or more employees, and with mean working hours of more than 15 hours per week. A total of 57% responded (N 5 2539), which is somewhat above average for surveys of this kind (Baruch and Holtom, 2008). The mean age was 43.79 years (SD 5 11.52), (range 19 to 66). The sample is representative of the working population in Norway, except for a minor overrepresentation of women (52% versus 47%; Hstmark and Lagerstrm, 2006).

Data were collected by a questionnaire measuring demographic variables, exposure to bullying, observed leadership behaviour of the respondents immediate superior, job satisfaction, subjective health complaints and various aspects of the psychosocial working environment. Only demographic variables and questions related to leadership behaviour are included in the present study. Leadership behaviour was measured using 22 items from the destructive leadership scale (Einarsen et al., 2002). Tyrannical leadership behaviour was measured using four items (Cronbachs alpha 5 0.70). Examples of items included has humiliated you, or other employees, if you/ they fail to live up to his/her standards and has spread incorrect information about you or your co-workers, in order to harm your/their position in the rm. Derailed leadership behaviour was measured by four items (Cronbachs alpha 5 0.71), examples of items being has used his/her position in the rm to prot nancially/materially at the companys expense and regards his/her sta more as competitors than as partners. Supportive disloyal leadership behaviour was measured by four items (Cronbachs alpha 5 0.65). Examples of items measuring this type of leadership behaviour are has behaved in a friendly manner by encouraging you/your co-workers to extend your/their lunch break and has encouraged you to enjoy extra privileges at the companys expense. Laissezfaire leadership behaviour was measured by four items (Cronbachs alpha 5 0.72) from the Multifactor Leadership Questionnaire (Bass and Avolio, 1990), an example being has avoided making decisions. To prevent response set among the participants, constructive leadership behaviour in the form of employee-centred, production-centred and change-centred leadership was included using six items from Ekvall and Arvonen (1991) (Cronbachs alpha 5 0.87), examples being gives recognition for good performance and encourages innovative thinking. Items measuring constructive leadership behaviour were distributed randomly among the items measuring destructive forms of leadership. Four response categories were employed (never, sometimes, quite often and very often or nearly always), and the respondents were asked to report on leadership behaviour
r 2009 British Academy of Management.

Destructive Leadership Behaviour which they had experienced during the last six months. To ensure the internal validity of the scales measuring the destructive forms of leadership, a series of exploratory factor analyses was conducted. The model that yielded the best t to the data was a ve-factor solution (w2 5 467.10; df 5 199; comparative t index (CFI) 5 0.95; goodness of t index (GFI) 5 0.88: consistent Akaike information criterion (CAIC) 5 933.01; root mean square error of approximation (RMSEA) 5 0.026) supporting the internal validity of the scale, which measures four destructive forms of leadership behaviour in addition to constructive leadership behaviour. Table 1 shows the t statistics for all the factor solutions. Data analysis Two methods are used to estimate the prevalence rate of destructive leadership: the operational classication method (OCM) and latent class cluster (LCC) analysis. The former denes a specic criterion that classies respondents as either exposed or not exposed to destructive leadership, based on their reports of their immediate superiors behaviour, a method commonly used in research on workplace bullying (e.g. Nielsen et al., 2009). As the denition of destructive leadership emphasizes repeated and systematic behaviour, the classication criterion employed was exposure to one or more types of destructive leadership behaviour during the last six months, quite often or very often or nearly always. Destructive leadership behaviour that is reported quite often or very often or nearly always is coded as 1, whereas all other frequencies are coded as 0. All instances are then added up and, when the sum is zero, the respondent is
Table 1. Fit statistics for the factor analyses solutions Model One factor: Leadership behaviour Two factor: Constructive and destructive leadership behaviour Three factor: Constructive, supportivedisloyal, destructive (laissez-faire, derailed and tyrannical) leadership behaviour Four factor: Constructive, supportivedisloyal, laissez-faire, destructive (derailed and tyrannical) leadership behaviour Five factor: Constructive, supportivedisloyal, laissez-faire, derailed, tyrannical leadership behaviour Satorra-Bentler scaled w2 3949.76 1654.05 821.56 572.40 467.10 df 209 208 206 203 199 RMSEA 0.093 0.058 0.038 0.030 0.026 CFI 0.81 0.88 0.93 0.94 0.95 GFI 0.52 0.64 0.81 0.82 0.88

443 not considered to be exposed to destructive leadership; otherwise, the respondent is considered to be exposed. Although a common method of reporting prevalence rates, some weaknesses have been pointed out concerning the OCM (Notelaers et al., 2006). First, the cut-o point provided by the OCM is an arbitrary choice that reduces a complex phenomenon to a simple eitheror phenomenon. Second, the number of items used may inuence the prevalence rate (Agervold, 2007). Third, subordinates who are frequently exposed to a wide range of specic destructive leadership behaviour, but where each specic behaviour only occurs sometimes, are not regarded as being exposed to destructive leadership. Of course, a low level of exposure to many dierent types of destructive leadership behaviour may still reect a systematic pattern in the leaders behaviour. To compensate for these potential weaknesses, we applied LCC analysis, which is a systematic way of classifying research subjects into homogeneous groups based on similarities in their responses to particular items, in our case the items describing the behaviour of their immediate supervisor. LCC analysis thus identies mutually exclusive groups based on the distribution of observations in an n-way contingency table of discrete variables (i.e. observed destructive leader behaviour). A goal of traditional LCC analysis is to determine the smallest number of latent classes, T, which is sucient to explain (account for) the associations observed between the manifest variables (the reported leadership behaviour) (Magidson and Vermunt, 2004). The analysis typically begins by tting the T 5 1 class (only one group of destructive leadership behaviour is reported) baseline model, which species mutual independence among the variables.

CAIC 4329.39 2042.32 1227.08 1003.80 933.01

r 2009 British Academy of Management.

444
Table 2. Means, standard deviations and inter-correlations for all continuous measures (N 5 2539) Variables 1 2 3 4 5 6 7 8 M SD 1 0.00 0.01 0.10** 0.04 0.10** 0.05* 0.01 2 0.05* 0.02 0.01 0.09** 0.03 0.03 3 4 5

M. S. Aasland et al.

Age 43.79 11.52 Number of employees in enterprise 279 835.9 Mean working hours per week 37.49 10.36 Tyrannical leadership 0.11 0.26 Derailed leadership 0.21 0.38 Supportivedisloyal leadership 0.29 0.38 Constructive leadership 1.44 0.66 Laissez-faire leadership 0.57 0.52

0.02 0.07** 0.08** 0.04* 0.02

0.60** 0.03 0.01 0.21** 0.29** 0.38** 0.54**

0.35** 0.08** 0.37**

*Correlation is signicant at the 0.05 level. **Correlation is signicant at the 0.01 level.

Assuming that this null model does not provide an adequate t to the data, a one-dimensional LCC model with T 5 2 classes (that distinguishes between destructive and non-destructive leadership behaviours) is then tted to the data. This process continues by tting successive LCC models to the data, increasing the number of classes each time thus implicitly introducing multidimensionality until the simplest model that provides an adequate t is found (Goodman, 1974; McCutcheon, 1987), and a model is found in which the latent variable can explain all of the associations among the reported behaviours (cf. Magidson and Vermunt, 2004). Dierent from traditional cluster methods (such as K-means clustering), LCC analysis is based on a statistical model that can be tested (Magidson and Vermunt, 2002a). In consequence, determining the number of latent classes is less arbitrary than when using traditional cluster methods (Notelaers et al., 2006). It can thus be seen as a probabilistic extension of Kmeans clustering (Magidson and Vermunt, 2002b). The LCC analysis will thus determine whether dierent groups exist among the respondents with respect to exposure to destructive leadership behaviour based on similarities in their response patterns (Notelaers et al., 2006). As the sample size in the present study is large, the level of signicance was set to po0.01.

Prevalence of destructive leadership behaviour using the OCM Destructive behaviour by superiors proved to be quite common, as 83.7% reported exposure to some kind of such behaviours. Yet, according to the operational criterion, 33.5% of the respondents reported exposure to at least one destructive leadership behaviour quite often or very often or nearly always during the last six months (Table 3). Employing this criterion, 21.2% were exposed to one or more instances of laissez-faire leadership behaviour, while 11.6% reported one or more instances of supportivedisloyal leadership behaviour. Furthermore, the prevalence of derailed leadership was 8.8%, with the prevalence rate of tyrannical leadership behaviour being 3.4% (Table 3).

LCC analysis concerning groups of targets The LCC analysis showed six separate clusters of reported leadership behaviour (see Table 4). For each of the six groups of respondents, Table 4 summarizes the prole output,1 retaining conditional probabilities by calculating the average conditional probability of each group responding never, sometimes, quite often or very often or nearly always to the items measuring the dierent forms of destructive leadership behaviour. The rst cluster, which we labelled no-destructiveness, included 39% of the respondents in the sample. This cluster is characterized by a high conditional probability of answering never to any of the items measuring the four kinds of
1 The prole output can be obtained by contacting the rst author of the paper.

Results
The inter-correlations, means and standard deviations for all the continuous measures used in the study are reported in Table 2.

r 2009 British Academy of Management.

Destructive Leadership Behaviour

Table 3. Exposure to constructive and destructive leadership behaviour quite often or more often Number of instances of behaviour exposed to 0 Constructive behaviour Tyrannical behaviour Derailed behaviour Supportivedisloyal behaviour Laissez-faire behaviour All destructive behaviourb 29.1 96.6 91.2 88.3 78.8 66.6 1 12.4 2.4 6.1 8.8 13.1 17.9 2 10.4 0.7 1.7 2.3 4.8 7.8 3 11.7 0.2 0.7 0.4 2.4 3.5

445

4 or morea 35.9 0.1 0.3 0.1 0.9 4.3

Frequency in per cent for rows (N 5 2539). a The constructive behaviour list consists of six items. The tyrannical, derailed, popular but disloyal lists each consist of four items. b All destructive behaviour is the sum of all tyrannical, derailed, supportivedisloyal and laissez-faire items (16 items).

Table 4. Average conditional probabilities (for all items concerned) expressed as percentages Cluster 1 Nondestructiveness Size Never Sometimes Quite often Very often/nearly always 0.39 0.91 0.09 0 0 Cluster 2 Cluster 3 Laissez- Sometimes laissez-faire, faire sometimes supportive disloyal 0.19 0.75 0.21 0.04 0.01 0.17 0.665 0.29 0.04 0 Cluster 4 Sometimes destructive 0.11 0.57 0.36 0.06 0.01 Cluster 5 Supportive disloyal 0.10 0.78 0.17 0.04 0.01 Cluster 6 Highly abusive 0.06 0.42 0.33 0.16 0.09

destructive leadership behaviour. Hence, these subordinates claim not to observe any kind of destructive behaviour on the part of their immediate superior. A second cluster of respondents, comprising 19% of the respondents, was labelled laissezfaire given their higher conditional probability of reporting exposure to items measuring laissezfaire leadership and their low conditional probabilities of reporting any tyrannical, derailed or supportivedisloyal leadership behaviour. Almost every item associated with laissez-faire leadership is reported more frequently in this cluster compared with the no-destructiveness cluster: has avoided making decisions (62% sometimes and 12% quite often), is likely to be absent when needed (61% sometimes and 17% quite often), has avoided getting involved in your work (58% sometimes and 24% quite often) and has steered away from showing concern about results (43% sometimes and 6% quite often). The third cluster was labelled sometimes laissez-faire and sometimes supportivedisloyal, as these subordinates report that their superiors
r 2009 British Academy of Management.

show some level of both these two kinds of behaviour. This cluster contained 17% of the respondents. Nearly every item associated with laissez-faire leadership is reported less frequently in this cluster than in the laissez-faire cluster. However, these respondents have a high probability of answering sometimes to items measuring both laissez-faire and supportivedisloyal leadership behaviour. Examples of items with a high average probability of being reported sometimes by these respondents are has behaved in a friendly manner by encouraging you/your coworkers to extend your/their lunch break (68% sometimes, 15% quite often), has encouraged you, or your co-workers, to take extra coee/ cigarette breaks as a reward for good work eort (61% sometimes, 8% quite often), is likely to be absent when needed (60% sometimes and 11% quite often), has avoided making decisions (56% sometimes and 6% quite often) and has avoided getting involved in your work (52% sometimes and 6% quite often). The probability of these respondents reporting any tyrannical or derailed leadership behaviour is very low.

446 The fourth cluster, which was labelled sometimes destructive, comprised 11% of the respondents. These respondents face sometimes destructive leadership behaviour in a variety of forms, be it laissez-faire, tyrannical or derailed leadership. Items measuring laissez-faire leadership behaviour are more frequently reported than in the sometimes laissez-faire and sometimes supportivedisloyal cluster, and items measuring tyrannical and derailed leadership behaviour are also more frequently reported, examples being has humiliated you, or other employees, if you/ they fail to live up to his/her standards (45% sometimes, 7% quite often), regards his/her sta more as competitors than as collaborators (43% sometimes, 10% quite often), has made it more dicult for you to express your views at a meeting, either by giving you too little speaking time or by placing you last (37% sometimes, 6% quite often) and has imitated or made faces (e.g. rolled eyes, stuck out tongue etc.) at you or other employees to show that he/she is not satised with your/others work eorts (36% sometimes, 7% quite often). The fth cluster of respondents, comprising 10% of the respondents, only report exposure to supportivedisloyal leadership behaviour. Nearly every item associated with supportive disloyal leadership is reported more frequently in this cluster than in the sometimes laissez-faire and sometimes supportivedisloyal cluster: has encouraged you, or your co-workers, to take extra coee/cigarette breaks as a reward for good work eort (60% sometimes, 27% quite often), has behaved in a friendly manner by encouraging you/your co-workers to extend your/their lunch break (60% sometimes, 26% quite often), has encouraged you to enjoy extra privileges at the companys expense (35% sometimes, 10% quite often) and has encouraged you or your co-workers to do private work/run private errands during working hours (35% sometimes, 4% quite often). The sixth cluster was labelled highly abusive and comprised 6% of the respondents in the study. Respondents in this cluster report high exposure to laissez-faire, tyrannical and derailed leadership behaviour. Their response patterns are characterized by a higher average probability of reporting that such behaviour occurs quite often and/or very often or nearly always than all the other clusters. Again, laissez-faire leadership

M. S. Aasland et al. behaviour seems to dominate, but respondents also report excessive exposure to tyrannical and derailed leadership behaviour. Examples of items that are frequently reported by respondents in this cluster are has avoided getting involved in your work (40% quite often, 23% very often or nearly always), is likely to be absent when needed (35% quite often, 29% very often or nearly always), attributes the companys success to his/her own abilities rather than the abilities of the employees (29% quite often, 19% very often or nearly always), has avoided making decisions (34% quite often, 22% very often or nearly always), regards his/her sta more as competitors than as collaborators (27% quite often, 12% very often or nearly always), has made it more dicult for you to express your views at a meeting, either by giving you too little speaking time or by placing you last (19% quite often, 11% very often or nearly always), has steered away from showing concern about results (19% quite often, 11% very often or nearly always) and has humiliated you, or other employees, if you/they fail to live up to his/her standards (15% quite often, 6% very often or nearly always). The LCC analysis shows that as many as 61% of all subordinates experience systematic exposure to varying forms and combinations of destructive leadership, with laissezfaire leadership again being the dominant form. Laissez-faire leadership may exist in isolation with either a low or a high frequency, or it may occur in combination with either supportive disloyal leadership or tyrannical and derailed leadership, again in more or less intense form.

Discussion
The present study shows that destructive leadership behaviour is very common. Depending on the estimation method, between 33.5% and 61% of all respondents report their immediate superiors as showing some kind of consistent and frequent destructive leadership during the last six months, while only about 40% report no exposure whatsoever to such leadership behaviour. Thus, destructive leadership behaviour is not a low base-rate phenomenon, but something that most subordinates will probably experience during their working life. The observed prevalence rates are far higher than those reported by
r 2009 British Academy of Management.

Destructive Leadership Behaviour Schat, Frone and Kelloway (2006) and Hubert and van Veldhoven (2001). Schat, Frone and Kelloway (2006) found that, during the last 12 months, 13.5% of the respondents were exposed to aggression from their superior, and, likewise, Hubert and van Veldhoven (2001) found an average prevalence rate of about 11%. However, these two studies only investigated one form of destructive leadership behaviour, namely aggressive behaviour. Consequently, they did not measure passive forms of destructive leadership nor destructive forms of leadership that may also include constructive elements. Furthermore, these studies did not measure destructive behaviour targeting the organization. Hence, measuring a broader range of destructive leadership behaviour, as was the case in the present study, will probably yield higher prevalence rates than measuring a narrower range of behaviour. Participants in this study reported exposure to all four forms of destructive leadership behaviour described in the presented conceptual model. While some 3.5% had been exposed to tyrannical leadership behaviour during the last six months, some 9% reported exposure to derailed leadership behaviour. About 11.5% had been exposed to supportivedisloyal leadership behaviour. Laissez-faire leadership behaviour had the highest prevalence rate at 21.2%, all according to the OCM. As, to our knowledge, no other studies have investigated the prevalence of tyrannical leadership behaviour as operationalized in the present study, it is dicult to evaluate whether a prevalence rate of about 3.5% should be considered high, average or low. However, if we look at the item level, this form of leadership behaviour overlaps with behaviour classied as abusive (Bies and Tripp, 1998) or aggressive leadership (Hubert and van Veldhoven, 2001; Schat, Frone and Kelloway, 2006). Thus, one can compare the prevalence rate for tyrannical leadership behaviour with the prevalence rates obtained in the study by Schat, Frone and Kelloway (2006) and the study by Hubert and van Veldhoven (2001). Although Schat, Frone and Kelloway found a prevalence rate of 13.5% for people who reported some kind of exposure to aggressive behaviour from their superior, only 5.5% reported frequent exposure to such behaviour. Furthermore, Schat, Frone and Kelloway measured exposure to aggressive behaviour durr 2009 British Academy of Management.

447 ing the last 12 months, while the present study has a time frame of only six months. In the study by Hubert and van Veldhoven (2001), only 2.1% of the respondents reported frequent (often or always) exposure to aggression from their leader. Thus, it can be concluded that the prevalence rate of tyrannical leadership behaviour in the present study is comparable to these studies. About 9% of the respondents reported that their immediate leader displayed derailed leadership behaviour, which, in addition to anti-subordinate behaviour, also involves anti-organizational behaviour, potentially making these leaders more visible to upper management, compared with leaders who display tyrannical leadership behaviour. The prevalence rate of exposure to this form of leadership behaviour could therefore be expected to be lower than for tyrannical leadership behaviour. That was not the case, however. Again the explanation may be that we measure leadership behaviour over a six-month period, as the potential positive eect tyrannical leadership behaviour may have on the organizations goal attainment is claimed to decline, and eventually disappear, over time (Ma, Karri and Chittipeddi, 2004), eventually developing into derailed leadership behaviour. Furthermore, experiencing antisubordinate behaviour such as bullying, humiliation or harassment may have such a profound eect on some subordinates that organization-oriented behaviour, be it constructive or destructive, is of minor importance to them. This assumption is supported by Baumeister and colleagues (2001), who found that bad experiences have a much stronger impact than good ones, resulting in a horns eect, whereby the leader is only perceived in negative terms. Supportivedisloyal leadership behaviour had a prevalence rate of about 11.5% according to the OCM, and it was the second most prevalent form of destructive leadership in the present study. This form of leadership behaviour involves behaving in a comradely and supportive manner towards subordinates, while at the same time violating the legitimate interests of the organization by, for example, working towards alternative goals to those of the organization, stealing from the organization or encouraging subordinates to engage in such activities (Einarsen, Aasland and Skogstad, 2007). To our knowledge, this form of destructive leadership behaviour has not been

448 investigated in earlier studies on destructive leadership. However, a recent qualitative, crosscultural study focusing on observations of senior leaders wrongdoing in ten dierent countries (Birkland, Glomb and Ones, 2008) showed that anti-organizational behaviour was reported more often than anti-subordinate behaviour, supporting the high prevalence rates of both derailed and supportivedisloyal leadership in the present study. Moreover, this form of destructive leadership may be underestimated in the present study as leaders at dierent levels in an organization have a continual opportunity to misuse their power and position for personal gain (Conger, 1990), relatively often without being noticed by subordinates. With such extensive opportunities to conceal these forms of destructive behaviour, we believe that a prevalence rate of some 11.5% should be considered a conservative estimate and, as such, as representing a serious problem for organizations. In the present study, laissez-faire leadership behaviour showed the highest prevalence rate, with 21.2% of the respondents reporting that they experienced at least one such leadership behaviour quite often or very often or nearly always during the last six months. Lack of adequate leadership, which is the case for laissez-faire leadership, is passive behaviour that may be harder to detect and intervene against than more active forms of destructive leadership behaviour, and it may also be more easily tolerated by senior managers (Skogstad et al., 2007). However, according to the LCC analysis, destructive leadership is not an eitheror phenomenon, but more a compound phenomenon with respect to both its nature and frequency. Based on this analysis, laissez-faire leadership behaviour is reported in three dierent forms, either as a stand-alone phenomenon, in combination with supportive disloyal leadership behaviour or as part of a compound phenomenon in which respondents report exposure to both tyrannical and derailed leadership behaviour as well as to laissez-faire leadership. Nineteen per cent of the respondents reported some laissez-faire leadership behaviour according to this analysis, while an additional 17% were exposed to a low-frequency combination of laissez-faire and supportivedisloyal leadership behaviour. Moreover, 11% of the respondents reported being exposed to tyrannical, derailed and laissez-faire leadership behaviour at

M. S. Aasland et al. a medium frequency, while some 6% of the respondents reported high exposure to all these three destructive forms of leadership. Hence, laissez-faire leadership seems to be much more prevalent than indicated by the use of the OCM. The present study focuses on exposure to destructive leadership behaviour during a sixmonth period. Hence, the behaviours reported in the dierent clusters are not necessarily experienced simultaneously. Laissez-faire leadership behaviour may over time evolve into more active forms of destructive leadership behaviour, in the same way that tyrannical leadership behaviour may evolve into derailed leadership behaviour over time (Ma, Karri and Chittipeddi, 2004). However, longitudinal studies are needed to investigate whether such an escalation process is involved. According to the LCC analysis, about 17% of the respondents reported consistent and systematic exposure to supportivedisloyal leadership behaviour in combination with laissez-faire leadership behaviour. Moreover, 10% reported more frequent exposure to supportivedisloyal leadership behaviour alone. Hence, low-frequency supportivedisloyal leadership behaviour seems fairly common in Norwegian working life. Supportivedisloyal leadership behaviour may even be perceived as just being nice, and not as constituting destructive behaviour. However, over time, even low-frequency supportivedisloyal leadership may entail considerable expense for the organization. Additionally, our results show that highly abusive leaders display a range of dierent forms of behaviour, be it laissez-faire, derailed or tyrannical behaviour. A total of 17% of the respondents report such compound negative behaviour, even though only 6% report it happening on a very frequent basis. Thus, the results substantiate our assumption that destructive leadership is a phenomenon comprising dierent classes of behaviour. Focusing on aggression by a leader (Hubert and van Veldhoven, 2001; Schat, Frone and Kelloway, 2006) or abusive supervision only (Tepper, 2007) may therefore result in a too constricted focus, resulting in a partial picture of the complex phenomenon of destructive leadership. Furthermore, moderate correlations exist between the dierent forms of destructive leadership behaviour and constructive leadership behaviour (Table 2), indicating that destructive leadership is not something that exists apart from constructive leaderr 2009 British Academy of Management.

Destructive Leadership Behaviour ship but is an integrated part of the behavioural repertoire of leaders. A similar conclusion can be drawn from an English study on harassment by managers (Rayner and Cooper, 2003) in which, of 72 managers evaluated by subordinates, only 11.1% were exclusively perceived as being a tough manager, with 9.7% of the managers being exclusively perceived as angels. In the same study, about 28% of the managers were evaluated as either tough managers with supporters, middlers with victims or as angels with victims. Thus, leaders behave both destructively and constructively, demonstrating dierent behaviour and dierent combinations of behaviour in relation to dierent subordinates, who again may react dierently (Rayner and Cooper, 2003). This has implications for leadership research, which, in the past, has mainly seen leadership as being either constructive or destructive, with the latter being an anomaly. Furthermore, experiencing both constructive and destructive behaviour from the same source may have more distressing eects on the target than being exposed to destructive behaviour alone (Duy, Ganster and Pagon, 2002; Major et al., 1997). Methodological issues A notable strength of the present study is its use of a representative sample of the Norwegian workforce (Hstmark and Lagerstrm, 2006). Previous research in this eld has mainly been carried out in small non-representative samples, giving rise to uncertainty about the generalizability of the ndings. Furthermore, the present study employs two methods of calculating prevalence rates. While the OCM implicitly assumes that only two groups of respondents exist (those exposed and those not exposed), the LCC method identies several dierent groups with respect to observations of destructive leadership in ones immediate superior (Notelaers et al., 2006), taking both the nature and the frequency of the behaviour into account. Research on bullying in the workplace shows that the LCC method displays better construct and predictive validity than the OCM (Notelaers et al., 2006). Thus, the LCC method seems to result in a more convincing and comprehensive estimate of the prevalence of destructive leadership. A limitation of the present study that should be mentioned is that the Cronbachs alpha of the
r 2009 British Academy of Management.

449 supportivedisloyal leadership behaviour subscale of the destructive leadership scale is low, thus questioning the internal consistency of this subscale (Hinkin, 1998). Consequently, the results obtained using this subscale should be interpreted with caution, and the subscale should be investigated further in future studies. The present study is based on subordinates appraisal. Phillips and Lord (1986) distinguish between two types of accuracy in such leader appraisals, namely classication accuracy and behavioural accuracy. Classication accuracy refers to the ability to adequately classify a persons behaviour, while behavioural accuracy is the ability to detect the presence of a specic type of behaviour (Bretz, Milkovich and Read, 1992). If classication accuracy is lacking, subordinates may be inuenced by a halo (Thorndike, 1920) or horns eect (Bligh et al., 2007). The halo eect consists of an overall positive evaluation of a leader based, for example, on a single positive characteristic or action (Thorndike, 1920), while the horns eect refers to the reverse situation in which an overall negative appraisal is made based on one salient failure or negative characteristic (Bligh et al., 2007), which may prevent subordinates from noticing the variety and nuances in their leaders behaviour. However, Cardy and Dobbins (1994) emphasize that subordinates are valid sources of information, as they are often well placed to observe and evaluate their leaders performance and leadership style. Moreover, as negative experiences and information are processed more thoroughly than are positive information and experiences (Baumeister et al., 2001), subordinates appraisals of destructive leadership behaviour probably have high behavioural accuracy.

Conclusion
Destructive forms of leadership behaviour are highly prevalent, at least in their less severe forms, including the passive form of laissez-faire leadership. Considering the negative eects of destructive leadership for both subordinates and the organization documented in several studies (Bamberger and Bacharach, 2006; Mitchell and Ambrose, 2007; Tepper, 2000), destructive leadership constitutes a serious problem in contemporary working life. Furthermore, destructive

450 leadership behaviour comes in many shapes and forms, categorized along two basic dimensions, namely pro-organizational versus anti-organizational behaviour and pro-subordinate versus anti-subordinate behaviour, meaning that a leader over a period of time may display constructive as well as destructive behaviour. A high prevalence rate of destructive leadership behaviour has important implications for theory development regarding destructive leadership in particular, but also for leadership research in general. Leaders who behave in a destructive manner are not exceptional, nor can they be referred to as a few deviants, at least not as experienced by their subordinates. Moreover, destructive leadership behaviour is not a phenomenon that exists apart from constructive leadership, but must be viewed as an integral part of what constitutes leadership behaviour. Including this dark side of leadership, a more accurate and nuanced understanding of the very phenomenon of leadership behaviour may emerge, which in turn may contribute to the general understanding of both the nature and eectiveness of leadership, and to the development and management of leaders (Burke, 2006). Leaders may behave destructively for a variety of reasons, be it their personality, incompetence, perceived injustice or threat to their identity, nancial reasons, low organizational identication etc. Future studies should investigate the antecedents of the dierent forms of destructive leadership behaviour identied in this paper, as such knowledge could help us prevent such behaviour among leaders and develop tools for organizational reactions and the rehabilitation of leaders who act in breach of the legitimate interest of the organization.

M. S. Aasland et al.
Johnson and J. D. Douglas (eds), Crime at the Top: Deviance in Business and the Professions, pp. 90124. Philadelphia, PA: J. B. Lippincott. Aryee, S., Z. X. Chen, L. Sun and Y. A. Debrah (2007). Antecedents and outcomes of abusive supervision: test of a trickle-down model, Journal of Applied Psychology, 92, pp. 191201. Ashforth, B. (1994). Petty tyranny in organizations, Human Relations, 47, pp. 755778. Avolio, B. J. (1999). Full Leadership Development. Building the Vital Forces in Organizations. London: Sage. Bamberger, P. A. and S. B. Bacharach (2006). Abusive supervision and subordinate problem drinking: taking resistance, stress, and subordinate personality into account, Human Relations, 59, pp. 130. Baruch, Y. and B. C. Holtom (2008). Survey response rate levels in organizational research, Human Relations, 61, pp. 11391160. Bass, B. M. (1990a). Authoritarianism, power orientation, Machiavellianism, and leadership. In B. M. Bass (ed.), Bass & Stogdills Handbook of Leadership. Theory, Research, and Managerial Applications, pp. 124139. New York: Free Press. Bass, B. M. (1990b). Laissez-faire leadership versus motivation to manage. In B. M. Bass (ed.), Bass & Stogdills Handbook of Leadership. Theory, Research and Managerial Applications, pp. 544559. New York: Free Press. Bass, B. M. and B. J. Avolio (1990). Transformational Leadership Development: Manual for the Multifactor Leadership Questionnaire. Palo Alto, CA: Consulting Psychologists Press. Bass, B. M. and R. E. Riggio (2006). Transformational Leadership. London: Lawrence Erlbaum. Bass, B. M., B. J. Avolio, D. I. Jung and Y. Berson (2003). Predicting unit performance by assessing transformational and transactional leadership, Journal of Applied Psychology, 88, pp. 207218. Baumeister, R. F., E. Bratslavsky, C. Finkenauer and K. D. Vohs (2001). Bad is stronger than good, Review of General Psychology, 5, pp. 323370. Bies, R. J. and T. M. Tripp (1998). Two faces of the powerless. Coping with tyranny in organizations. In R. M. Kramer and M. A. Neale (eds), Power and Inuence in Organizations, pp. 203219. London: Sage. Birkland, A., T. M. Glomb and D. S. Ones (2008). A crosscultural comparison of senior leader wrongdoing. Paper presented at the 23rd Annual Conference of the Society for Industrial and Organizational Psychology, 11 April, San Francisco, CA. Blake, R. R. and J. S. Mouton (1985). The Managerial Grid III. Houston, TX: Gulf Publishing. Bligh, M. C., J. C. Kohles, C. L. Pearce, J. E. G. Justin and J. F. Stovall (2007). When the romance is over: follower perspectives of aversive leadership, Applied Psychology: An International Review, 56, pp. 528557. Bretz, R. D. J., G. T. Milkovich and W. Read (1992). The current state of performance appraisal research and practice: concerns, directions, and implications, Journal of Management, 18, pp. 321352. Burke, R. J. (2006). Why leaders fail. Exploring the dark side. In R. J. Burke and C. L. Cooper (eds), Inspiring Leaders, pp. 239248. London: Routledge. Buss, A. H. (1961). The Psychology of Aggression. London: Wiley.

References
Aasland, M. S., A. Skogstad and S. Einarsen (2008). The dark side of leadership: dening and describing destructive forms of leadership behaviour, Organizations and People, 15, pp. 2028. Adorno, T. W., E. Frenkel-Brunswik, D. J. Levinson and R. N. Sanford (1950). The Authoritarian Personality. New York: Harper. Agervold, M. (2007). Bullying at work: a discussion of denitions and prevalence, based on an empirical study, Scandinavian Journal of Psychology, 48, pp. 161172. Altheide, D. L., P. A. Adler, P. Adler and D. A. Altheide (1978). The social meaning of employee theft. In J. M.

r 2009 British Academy of Management.

Destructive Leadership Behaviour

Cardy, R. L. and G. H. Dobbins (1994). Performance Appraisal: Alternative Perspective. Cincinnati, OH: South-Western Publishing. Christie, R. and F. L. Geis (1970). Studies in Machiavellianism. New York: Academic Press. Conger, J. A. (1990). The dark side of leadership, Organizational Dynamics, 19, pp. 4455. Conger, J. A. and R. N. Kanungo (1987). Toward a behavioral theory of charismatic leadership in organizational settings, Academy of Management Review, 12, pp. 637647. Conger, J. A. and R. N. Kanungo (1988). The empowerment process: integrating theory and practice, Academy of Management Review, 13, pp. 471482. Ditton, J. (1977). Part-time Crime: An Ethnography of Fiddling and Pilferage. London: Billings. Duy, M. K., D. C. Ganster and M. Pagon (2002). Social undermining in the workplace, Academy of Management Journal, 45, pp. 331351. Einarsen, S., M. S. Aasland and A. Skogstad (2007). Destructive leadership behaviour: a denition and a conceptual model, Leadership Quarterly, 18, pp. 207216. Einarsen, S., A. Skogstad, M. S. Aasland and A. M. S. B. rsaker og konsekLseth (2002). Destruktivt lederskap: a venser [Destructive leadership: predictors and conse quences]. In A. Skogstad and S. Einarsen (eds), Ledelse pa Godt og Vondt. Eektivitet og Trivsel [Leadership for Better or Worse. Eciency and Job Satisfaction], pp. 233254. Bergen: Fagbokforlaget. Ekvall, G. and J. Arvonen (1991). Change-centered leadership: an extension of the two-dimensional model, Scandinavian Journal of Management, 7, pp. 1726. Fox, S. and P. E. Spector (1999). A model of work frustration aggression, Journal of Organizational Behavior, 20, pp. 915 931. Frischer, J. and K. Larsson (2000). Laissez-faire in research education an inquiry into a Swedish doctoral program, Higher Education Policy, 13, pp. 131155. Goodman, L. A. (1974). The analysis of systems of qualitative variables when some of the variables are unobservable. Part 1: A modied latent structure approach, American Journal of Sociology, 79, pp. 11791259. Hinkin, T. R. (1998). A brief tutorial on the development of measures for use in survey questionnaires, Organizational Research Methods, 1, pp. 104121. Hinkin, T.R and C. A. Schriesheim (2008). An examination of nonleadership. From laissez-faire leadership to leader reward omission and punishment omission, Journal of Applied Psychology, 93, pp. 12341248. Hogan, R. (1994). Trouble at the top: causes and consequences of managerial incompetence, Consulting Psychology Journal, 46, pp. 915. Hogan, R., R. Raskin and D. Fazzini (1990). The dark side of charisma. In K. E. Clark and M. B. Clark (eds), Measures of Leadership, pp. 343354. West Orange, NJ: Leadership Library of America. Hstmark, M. and B. O. Lagerstrm (2006). Underskelse om arbeidsmilj: destruktiv atferd i arbeidslivet [A study of work environment: destructive behaviours in working life]. /Statistics Dokumentasjonsrapport 44, Statistisk sentralbyra Norway, Oslo. House, R. J. and J. M. Howell (1992). Personality and charismatic leadership, Leadership Quarterly, 3, pp. 81108.

451
Hubert, A. B. and M. van Veldhoven (2001). Risk sectors for undesirable behaviour and mobbing, European Journal of Work and Organizational Psychology, 10, pp. 415424. Kellerman, B. (2004). Bad Leadership. What it is, How it Happens, Why it Matters. Boston, MA: Harvard Business School Press. Kelloway, E. K., J. Mullen and L. Francis (2006). Divergent eects of transformational and passive leadership on employee safety, Journal of Occupational Health Psychology, 11, pp. 7686. Kelloway, E. K., N. Sivanathan, L. Francis and J. Barling (2005). Poor leadership. In J. Barling, E. K. Kelloway and M. R. Frone (eds), Handbook of Work Stress, pp. 89112. Thousand Oaks, CA: Sage. Kets de Vries, F. R. and D. Miller (1985). Narcissism and leadership: an object relations perspective, Human Relations, 38, pp. 583601. Kipnis, D., S. Schmidt, K. Price and C. Stitt (1981). Why do I like thee: is it your performance or my orders?, Journal of Applied Psychology, 66, pp. 324328. Lewin, K., R. Lippitt and R. K. White (1939). Patterns of aggressive behaviour in experimentally created social climates, Journal of Social Psychology, 10, pp. 271301. Lipman-Blumen, J. (2005). The Allure of Toxic Leaders. Why We Follow Destructive Bosses and Corrupt Politicians and How We Can Survive Them. Oxford: Oxford University Press. Lubit, R. (2004). The tyranny of toxic managers: applying emotional intelligence to deal with dicult personalities, Ivey Business Journal, 68, pp. 17. Ma, H., R. Karri and K. Chittipeddi (2004). The paradox of managerial tyranny, Business Horizons, 4, pp. 3340. Magidson, J. and J. K. Vermunt (2002a). Latent class modeling as a probabilistic extension of K-means clustering, Quirks Marketing Research Review, 20, pp. 7780. Magidson, J. and J. K. Vermunt (2002b). Latent class models for clustering: a comparison with K-means, Canadian Journal of Marketing Research, 20, pp. 3744. Magidson, J. and J. K. Vermunt (2004). Latent class models. In D. Kaplan (ed.), The SAGE Handbook of Quantitative Methodology for the Social Sciences, pp. 175198. Thousand Oaks, CA: Sage. Major, B., J. M. Zubek, M. L. Cooper, C. Cozzarelli and C. Richards (1997). Mixed messages: implications of social conict and social support within close relationships for adjustment to a stressful life event, Journal of Personality and Social Psychology, 72, pp. 13491363. McCall, M. W. J. and M. M. Lombardo (1983). O the Track: Why and How Successful Executives Get Derailed. Technical Report No. 21. Greensboro, NC: Center for Creative Leadership. McCutcheon, A. (1987). Latent Class Analysis. Newbury Park, CA: Sage. Mitchell, M. S. and M. L. Ambrose (2007). Abusive supervision and workplace deviance and the moderating eects of negative reciprocity beliefs, Journal of Applied Psychology, 92, pp. 11591168. Nielsen, M. B., A. Skogstad, S. B. Matthiesen, L. Glas, M. S. Aasland, G. Notelaers and S. Einarsen (2009). Prevalence of workplace bullying in Norway: comparisons across time and estimation methods, European Journal of Work and Organizational Psychology, 18, pp. 81101.

r 2009 British Academy of Management.

452
Notelaers, G., S. Einarsen, H. De Witte and J. K. Vermunt (2006). Measuring exposure to bullying at work: the validity and advantages of the latent class cluster approach, Work and Stress, 20, pp. 288301. Padilla, A., R. Hogan and R. B. Kaiser (2007). The toxic triangle: destructive leaders, susceptible followers, and conducive environments, Leadership Quarterly, 18, pp. 176194. Phillips, J. S. and R. G. Lord (1986). Notes of the practical and theoretical consequences of implicit leadership measurement, Journal of Management, 12, pp. 3141. Rayner, C. and C. L. Cooper (2003). The black hole in bullying at work research, International Journal of Management and Decision Making, 4, pp. 4764. Sackett, P. R. and C. J. DeVore (2001). Counterproductive behaviors at work. In N. Anderson, D. S. Ones, H. K. Sinangil and C. Viswesvaran (eds), Handbook of Industrial, Work and Organizational Psychology, Vol. 1, pp. 145164. London: Sage. Schat, A. C. H., M. R. Frone and E. K. Kelloway (2006). Prevalence of workplace aggression in the U.S. workforce: ndings from a national study. In E. K. Kelloway, J. Barling and J. J. Hurrell (eds), Handbook of Workplace Violence, pp. 4790. Thousand Oaks, CA: Sage. Shackleton, V. (1995). Leaders who derail. In V. Shackleton (ed.), Business Leadership, pp. 89100. London: Thomson.

M. S. Aasland et al.
Skogstad, A., S. Einarsen, T. Torsheim, M. S. Aasland and H. Hetland (2007). The destructiveness of laissez-faire leadership behavior, Journal of Occupational Health Psychology, 12, pp. 8092. Tepper, B. J. (2000). Consequences of abusive supervision, Academy of Management Journal, 43, pp. 178190. Tepper, B. J. (2007). Abusive supervision in work organizations: review, synthesis, and research agenda, Journal of Management, 33, pp. 261289. Thorndike, E. L. (1920). A constant error on psychological rating, Journal of Applied Psychology, 4, pp. 2529. Vredenburgh, D. and Y. Brender (1998). The hierarchical abuse of power in work organizations, Journal of Business Ethics, 17, pp. 13371347. Wilson, B. (1991). U.S. businesses suer from workplace trauma. Heres how to protect the mental health of your work force and the scal health of your company, Personnel Journal, 70, pp. 4750. Zapf, D., S. Einarsen, H. Hoel and M. Vartia (2003). Empirical ndings on bullying in the workplace. In S. Einarsen, H. Hoel, D. Zapf and C. L. Cooper (eds), Bullying and Emotional Abuse in the Workplace. International Perspectives in Research and Practice, pp. 103126. London: Taylor & Francis. Zellars, K. L., B. J. Tepper and M. K. Duy (2002). Abusive supervision and subordinates organizational citizenship behavior, Journal of Applied Psychology, 87, pp. 10681076.

Merethe Schanke Aasland is a licensed clinical psychologist, currently working as a research fellow and a PhD student at the University of Bergen completing her PhD thesis on destructive leadership in organizations. She is also a member of the Bergen Bullying Research Group. Main research topics are leadership, counterproductive behaviours and workplace bullying. Dr Anders Skogstad is a Professor at the Faculty of Psychology, University of Bergen, and member of the Bergen Bullying Research Group. He is a specialist in work and organizational psychology, Norwegian Psychological Association. His main research topics are destructive leadership in organizations, psychosocial factors at work, counterproductive behaviours and workplace bullying. Guy Notelaers is currently funded by the Norwegian Research Council to deliver a PhD on a stressoriented approach to explain workplace bullying, and is a member of the Bergen Bullying Research Group. His main research topics are methods and statistics in the area of work and organizational research. Morten Birkeland Nielsen has a PhD in psychology and is currently employed as a postdoctoral research associate at the University of Bergen. He is a member of the Operational Psychology Research Group and the Bergen Bullying Research Group. His main research topics are workplace bullying, destructive leadership, whistle-blowing and research methodology. le Einarsen is a Professor at the Faculty of Psychology, University of Bergen, and managing Dr Sta director of the Bergen Bullying Research Group. His main research topics are destructive leadership in organizations, creativity and innovation, workplace bullying and sexual harassment.

r 2009 British Academy of Management.

Copyright of British Journal of Management is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.