a b s t r a c t
This study proposes a process model of the antecedents of both constructive and destructive
leadership. As task difficulty increases, a leader’s limited attentional resource capacity may become
overwhelmed by the experience of high levels of negative emotions, resulting in self-regulating
impairment and destructive leadership. When task difficulty is low, or when negative
emotions do not overwhelm attentional resource capacity, then self-regulation is effective, giving
rise to constructive leadership. We test our modelwith 161 leaders in the field and find good support
for our model in the prediction of transformational leadership and abusive supervision as
specific examples of constructive and destructive leadership.
For many years organizational scholars have focused predominantly on constructive forms of leadership, which encompass prosubordinate
and pro-organization leader behaviors (Aasland, Skogstad, Notelaers, Nielsen, & Einarsen, 2010; Schyns & Schilling,
2013). More recently, scholarly attention has shifted to anti-subordinate or destructive forms of leadership (Padilla, Hogan, &
Kaiser, 2007), such as petty tyranny (e.g., Kant, Skogstad, Torsheim, & Einarsen, 2013) and abusive supervision (e.g., Mawritz,
Folger, & Latham, 2014), which describe leaders who bully, harass or humiliate subordinates (Einarsen, Aasland, & Skogstad, 2007).
There are good reasons for this shift in focus, asmounting evidence suggests that destructive leadership harms the mental and physical
health of employees and degrades organizational performance (Aasland et al., 2010; Hershcovis & Rafferty, 2012; Schyns &
Schilling, 2013). Given the individual and organizational harmattributed to destructive leadership, it is imperative thatwe better understandwhy
a leader might demonstrate destructive rather than constructive forms of behavior. Armedwith such knowledge, organizations
could implement strategies to reduce the prevalence and impact of destructive leadership.
The contemporary view holds that negative contextual factors like organizational injustice and abuse fromhigher-level managers
lead to destructive forms of leadership, for example, abusive supervision (Hershcovis & Rafferty, 2012; Tepper, 2007). In addition, a
number of leader dispositions such as anger, anxiety and poor self-regulation have been linked to destructive leadership (Kant
et al., 2013; Krasikova, Green, & LeBreton, 2013; Mawritz et al., 2014). In reviewing this evidence, Krasikova et al. (2013) argue
that destructive leadership is the product of both contextual and dispositional factors (see also Eubanks & Mumford, 2010a). In particular,
they propose that destructive leadership is a response to goal blockage when leaders lack sufficient psychological resources,
such as attention or emotional self-regulation (Krasikova et al., 2013). Goal blockage occurs when leaders experience difficulty
achieving their goals, however Krasikova et al. (2013) argue that the likelihood of displaying destructive versus constructive leadership
in response to goal blockage depends on a leader’s characteristics and contextual factors.
By incorporating goal blockage and psychological resources in their proposal, Krasikova et al. (2013) answer recent calls to enhance
leadership theory by integrating contextual and dispositional factors (see Avolio, 2007; Zaccaro, 2012). In addition, their
focus on goal blockage highlights the difficult and demanding nature of managerial work in modern organizations (Joosten, van
Dijke, Van Hiel, & De Cremer, 2014). Senior leaders in particular often face ill-defined and multi-faceted problems that threaten the
survival of the organizations they lead (Hambrick, 1989; Sherman, Hitt, DeMarie, & Keats, 1999). Furthermore, many operate in a
globally connected environment where deadlines are tight and vast quantities of information compete for their attention. Paradoxically,
decision-making becomes more difficult and stressful when more information is available, options for action increase and outcomes
are crucially important (Hambrick, Finkelstein, & Mooney, 2005; Miller & Cohen, 2001). Such dynamic and demanding work
environments have been associated with proactive forms of behavior (Parker, Bindl, &Wu, 2013; Parker,Williams, & Turner, 2006)
and transformational leadership (Bass, 1990b; Bass & Avolio, 1997). This raises an important question: In what way does difficult
or demanding work increase the chance of destructive leadership in some leaders but not others?
While Krasikova et al. (2013) describe some of the factors associated with destructive leadership, they do not explain the process
bywhich the depletion of psychological resources leads to destructive leadership (e.g., Dinh & Lord, 2012).We extend their proposal
in the present study to argue that constructive leadership is more likelywhen there is a sufficient level of psychological resources.We
suggest that reducing the adverse impact of destructive leadership starts with understanding the process by which psychological resources,
namely, attention and self-regulation, influence leader emotions and behavior. Such insight could inform the selection and
development of organizational leaders, particularly for demanding and stressful roles.
This paper has two key objectives: (1) to present a processmodel of self-regulation and leadership, particularly one that describes
howa self-regulatorymechanismis related to either transformational leadership (i.e., constructive leadership) or abusive supervision
(i.e., destructive leadership); and (2) to demonstrate howthis self-regulatory mechanismcan be operationalized and tested in a field
experiment utilizing common paper-and-pencil measures. Our central tenet is that sufficient attentional resource capacity is necessary
for effective self-regulation and transformational leadership during demanding performance tasks, whereas insufficient attentional
resource capacity leads to abusive supervision and heightened negative emotions in the same situation. We next make four
important points in relation to psychological resources (i.e., attention and self-regulation) that are central to our proposal.
First, we focus on attention and draw specifically from Beal,Weiss, Barros, and MacDermid (2005), who argue that attentional resources
serve as “an ‘engine’ specifically for the act of self-regulation” (p. 1058). Accordingly, we argue that effective self-regulation requires
sufficient attentional resource capacity. Second, we conceptualize self-regulation as a state-like construct, varying in strength
over time (e.g., Baumeister, Vohs, & Tice, 2007), and not as a trait-like or static construct (see Cervone, Shadel, Smith, & Fiori, 2006).
Third, the labels self-regulation and self-control are often used interchangeably; however, self-control is typically considered to be a
“deliberate, conscious, effortful subset of self-regulation” (Baumeister et al., 2007, p. 351). Self-control is typically defined as “exerting
control over one’s actions and inner states so as to bring them into line with meaningful standards such as goals, values and expectations”
(Bertrams, Englert,Dickhauser, & Baumeister, 2013, p. 669). However, we are primarily concernedwith the self-regulatory mechanism
that determines one’s capacity to exercise self-control (e.g., Cervone et al., 2006) and not a leader’s conscious or deliberate acts of
self-control (e.g., Tsui & Ashford, 1994). Finally, like others, we hold that self-regulation is highly adaptive and enables leaders to “engage
in goal-directed behavior to bring about long-term desirable outcomes” (Hagger,Wood, Stiff, & Chatzisarantis, 2010, p. 495).
We structure our introduction in the followingway. First,we briefly describe transformational leadership and abusive supervision,
which are examples of constructive and destructive leadership used in our study. Following Krasikova et al. (2013), we also examine
leader characteristics (i.e., attentional resources, self-regulation and negative emotions) and negative contextual factors as likely antecedents
of both forms of leadership. Next, we introduce the context-appropriate balanced attention model (CABA; MacCoon,
Wallace, & Newman, 2004) and explain howa neurocognitive (brain-based) attentionalmechanism regulates leader cognitions, emotions
and behavior. This model is well suited to our purpose because it explains the automatic process by which negative emotions
interfere with the allocation of limited-capacity attentional resources during a demanding performance task. We explain how this
process leads to self-regulatory impairment, heightened negative emotions and destructive leadership.
Theory and hypotheses development
The antecedents of transformational leadership and abusive supervision
Organizational leaders are often pressured to respond rapidly and accurately to work tasks and frequently make decisions based
on scarce or unreliable information. Such decision-making can be demanding and stressful, particularlywhen a leader’s decisions impact
followers and the organization. Effective leadership under such dynamic and demanding conditions requires leaders who are
proficient, adaptable and proactive (Griffin, Neal, & Parker, 2007). Transformational leadership is awell-established formof constructive
leadership often associated with proactive behavior during dynamic and demanding situations, for example, organizational
change or crisis (Bass, 1990b; Franke & Felfe, 2011; Podsakoff,MacKenzie,Moorman, & Fetter, 1990). Transformational leadersmotivate
others by providing themwith a value-laden vision, intellectual stimulation, inspirational communication, supportive leadership
and personal recognition (Rafferty & Griffin, 2004).
However, some leaders may be vulnerable to cognitive overload and stress in dynamic and demanding situations, demonstrating
few if any constructive leadership behaviors (Eubanks & Mumford, 2010b). Aggressive or hostile behavior could be one possible reaction
fromleaderswho feel threatened in such situations (Anderson & Bushman, 2002). This behavior reflects “abusive supervision,”
a form of destructive leadership describing leaders who “engage in the sustained display of hostile verbal and nonverbal behaviors,
excluding physical contact” (Tepper, 200, p. 178). Examples of abusive supervision include intimidation, withholding vital information,
and blaming or ridiculing a follower in front of others.
In a recent theoretical review, Krasikova et al. (2013) identified leader anger, anxiety and poor self-regulation as known antecedents
of destructive leadership. In addition, several studies have reported a link between abusive supervision and negative contextual
factors, such as organizational injustice, psychological contract violation, abuse fromhigher-level managers and aggressive organizational
norms (see Hershcovis & Rafferty, 2012;Mawritz et al., 2014; Tepper, 2007). Furthermore, somescholars argue that destructive
leadership is the product of both contextual and dispositional factors (Eubanks & Mumford, 2010a; Krasikova et al., 2013). Evidence
supporting this view can be found in two recent empirical studies. Joosten et al. (2014), following the ego depletion literature
(Baumeister, Bratslavsky,Muraven, & Tice, 1998), reported that demandingwork situationswere associatedwith unethical leader behavior
(i.e., destructive leadership) in resource-depleted leaders. Mawritz et al. (2014), in a study based on the cognitive theory of
stress (Lazarus & Folkman, 1984), found that supervisors’ hindrance stress and negative emotions (i.e., anxiety and anger) mediated
the relationship between “exceedingly difficult” job goals and follower-rated abusive supervision. That is, a leader’s perception of job
goals being more difficult led to higher stress, negative emotions and abusive supervision.
These two recent studies both draw on prominent psychological resource theories and provide preliminary evidence linking destructive
leadership to demanding work situations and a leader’s psychological resources. However, we could not find an equivalent
theory or empirical study examining the influence of contextual factors on a leader’s psychological resources in the more extensive
constructive leadership literature. Contextual factors have nevertheless been linked to transformational leadership, with several
scholars arguing that transformational leaders are more likely to emerge in “dynamic” rather than stable work environments (Bass,
1990a; Bass & Avolio, 1997; De Hoogh, Den Hartog, & Koopman, 2005). Effective performance in such uncertain work environments
requires individuals who are flexible and proactive (Griffin et al., 2007). Proactivity is generally associated with constructive forms of
leadership (see Crossley, Cooper, &Wernsing, 2013;Wu &Wang, 2011), including transformational leadership (Bass, 1990b; Bass &
Avolio, 1997). Furthermore,while the positive relationship between certain leader dispositions (i.e., emotional stability and conscientiousness;
Costa & McCrae, 1991) and transformational leadership has been widely cited (e.g., Bono & Judge, 2004), several studies
have reported contradictory or inconclusive findings, particularly when contextual factors vary, e.g., dynamic versus stable work environments
(De Hoogh et al., 2005; Judge & Bono, 2000; Lim & Ployhart, 2004; Ployhart, Lim, & Chan, 2001).
Thus it appears that the constructive leadership literature offers no clear and consistent reason as to why transformational leadership
ismore likely to occur than abusive supervision in dynamicwork environments. As previously argued, dynamic work environments
can be perceived as difficult and stressful, which constitutes a negative contextual factor associated with abusive supervision
(Hershcovis & Rafferty, 2012). In addition, some scholars have suggested that a leadermight display both constructive and destructive
leadership over a period of time (Aasland et al., 2010; Einarsen et al., 2007). Extending this idea,work environments that are perceived
to be dynamic, demanding or stressful can potentially impact leaders’ psychological resources such that they experience difficulty regulating
their cognitions, emotions and behavior (Baumeister et al., 2007; Krasikova et al., 2013). As previously argued, a sufficient level
of self-regulation is needed to pursue goal-directed behavior and achieve long-term individual and organizational outcomes
(e.g., Hagger et al., 2010; Yeow & Martin, 2013). This requirement suggests that effective self-regulation leads to proactive behavior
and transformational leadership, rather than abusive supervision.What remains unclear, however, is the process by which demanding
or stressful work environments impact psychological resources and leadership.
A partial explanation can be inferred froma study of leader performance that reported a positive correlation between trait anxiety
and performance for leaders with higher cognitive ability, while no correlation was found for those with lower cognitive ability
(Perkins & Corr, 2005). More generally, a substantial body of research attributes performance impairment to worrying thoughts
and anxiety interfering with attention and subsequently reducing the availability of cognitive resources for task-processing activities
(e.g., Bertrams et al., 2013; Eysenck & Calvo, 1992; Hardy, Beattie, &Woodman, 2007). Other scholars argue that anxiety interactswith
motivational constructs (e.g., extra effort, positive feedback) that compensate for the effect of anxiety and improve performance, but
more so on less difficult tasks (e.g., Eysenck, Derakshan, Santos, & Calvo, 2007; Humphreys & Revelle, 1984).
This suggests that the relationship between anxiety and performancemight depend on the difficulty of the task, such that anxiety
is associatedwith impaired performance on difficult tasks but results in unimpaired or even improved performance on easy tasks (see
also Smillie, Yeo, Furnham, & Jackson, 2006). It is also conceivable that anxiety is linked to impaired performance in situationswhere a
leader’s cognitive resources are restricted or limited, as reported in the study by Perkins and Corr (2005). Hence, it is possible that dynamic
or demanding work environments where leaders experience difficult or stressful performance tasks give rise to negative emotions
(e.g., anxiety and anger) that adversely impact their attentional resource capacity, leading to destructive forms of leadership.
Building on this idea, we next draw on a neurocognitive model of self-regulation, proposed by MacCoon et al. (2004), to explain
themechanismby which anxiety and other negative emotions interfere with attention, leading to a reduction in attentional resource
A neurocognitive model of self-regulation
The context-appropriate balanced attentionmodel (CABA;MacCoon et al., 2004) describes a neurocognitive limited-capacity “selective
attention” mechanismthat regulates cognitions, emotions and behavior. In thismodel selective attention is conceptualized as a
“top-down” self-regulatorymechanism responsible for enhancing or suppressing appropriate and inappropriate cognitions, emotions
or behaviors (see Botvinick, Braver, Barch, Carter, & Cohen, 2001; Posner & Rothbart, 1998; Posner, Rothbart, Sheese, & Tang, 2007).
Accordingly, particular cognitions, emotions and behaviors can be represented as networks of coactivated neurons that are activated
automatically in a “bottom-up” manner in response to particular stimuli. Difficult cognitive tasks, for example, draw on limited attentional
capacity, causing a reduction in available attentional resources for non-dominant stimuli (i.e., non-task-related). Furthermore,
selective top-down attention is attracted to the network that is currently most activated and only activates non-dominant networks
when attentional capacity is available. Top-down attention is also required to resolve network “coactivation,” that is, when dominant
and non-dominant networks compete for limited attentional capacity. This might occur, for example, when non-dominant worrying
thoughts and anxiety competewith dominant task-related networkswhen a leader presents poor business results to a hostile board of
The CABAmodel is relevant because it describes the seemingly automatic process bywhich threatening stimuli (i.e., negative contextual
factors) can deplete attentional resource capacity, leading to self-regulatory impairment.MacCoon et al. (2004) attribute selfregulatory
impairment to an “emotion-driven narrowing of attention,” arguing that as “capacity decreases, individuals will continue
to process bias-related cues (their priority) at the expense of processing cues unrelated to this bias” (p. 436). For example, the authors
suggest that highly threat-sensitive individuals should process neutral and threatening stimuli equally, when attentional capacity is
available, but should preferentially process threatening stimuli as attentional capacity decreases (see also Baskin-Sommers,Wallace,
MacCoon, Curtin, & Newman, 2010; Derryberry & Reed, 2002; Eysenck et al., 2007). As attentional capacity is increasingly allocated to
these threat networks, less capacity is available to activate those networks responsible for primary tasks (e.g., problem-solving), leading
to self-regulatory impairment and poor task performance. For example, a leader who feels angry and resentful when denied an
annual bonusmight unintentionally react abusivelywhen a subordinate challenges an unpopular leadership decision. This occurs because
the leader’s attentional resources are over-allocated to processing the negative emotions related to losing the bonus, leaving
less capacity available to respond appropriately to the subordinate’s actions.
The CABAmodel provides a parsimonious framework, supported by emerging neuroscientific evidence (e.g., the role of themedial
prefrontal cortex, see van Noordt & Segalowitz, 2012), thatwe use to integrate negative contextual factors, negative emotions and attentional
resources to predict either constructive or destructive leader behaviors. In addition, the limited-capacity self-regulatory
mechanismcentral to this model is conceptualized as a dynamic state-like construct. This viewalignswith Beal et al. (2005),who suggest
that regulatory resources “determine a person’s ability to control the allocation of their [cognitive] resources” (p. 1058, the terms
“regulatory resources” and “attentional resources” are synonymous). Beal et al. (2005) also argue that a temporal unit of performance,
or “performance episode,” is needed to link state-like constructs, such as emotions and attentional resources, to performance. A performance
episode is a relatively short, goal-directed task or behavior whereby, “performance during an episode is a joint function of resource
level and resource allocation” (Beal et al., 2005, p. 1057, original italics).
We believe that many leadership activities or tasks fit the description of a performance episode, including conducting an annual
performance appraisal, leading a team meeting or delivering a presentation. Importantly, we suggest that the level of task difficulty
influences the level and allocation of limited attentional resources necessary for effective self-regulation and subsequent leader emotions
and behavior. We next describe the process by which this occurs and how fluctuations in leaders’ self-regulation impact their
emotions and behavior over the course of demanding performance tasks.
Toward a process model of self-regulation and leadership
During the course of a demanding performance episode (i.e., one that is difficult, ambiguous or stressful)we propose that a leader’s
emotions and behaviorwill evolve in the followingway. First, difficult performance tasks place heavy demands on a leader’s cognitive
and attentional resources (Miller & Cohen, 2001; Shamosh & Gray, 2007; Shamosh et al., 2008). As attentional capacity decreases over
time due to heavy task-processing demands, threat-sensitive leaders will focus on processing threat-related negative thoughts
(Derryberry & Reed, 2002; MacCoon et al., 2004). Attending to these negative thoughts increases the strength of negative emotions,
which further deplete attentional resources. The corresponding reduction in attentional capacity degrades task performance, increases
the likelihood of errors, initiates more threat-related negative thoughts, and further increases the activation and allocation
of attentional resources to these threat networks (Miller & Cohen, 2001; Shamosh et al., 2008). As a limited-capacity construct, the
CABA model predicts that this ongoing processwill result in diminishing levels of task-related attentional resources, increasing levels
of negative emotion and a weakening or failure in self-regulation.
Accordingly, we propose a moderated mediation model predicting leader behavior over the course of a demanding performance
task (to be clear, we do not expect the same effect for an easy or simple task). Specifically, we propose that attentional resource capacitymoderates
the relationship between pre-task negativeemotions and self-regulation,while self-regulationmediates the relationship
between pre- and post-task negative emotions. Furthermore, higher levels of attentional capacitywill reduce the adverse effect of
pre-task negative emotions on self-regulation, leading to effective self-regulation, lower post-task negativeemotions and constructive
leader behavior. Conversely, lower levels of attentional capacitywill increase the adverse effect of pre-task negative emotions, leading
to ineffective self-regulation, higher post-task negative emotions and destructive leader behavior (see Fig. 1).
Effective self-regulation during a demanding performance task therefore depends on the efficient allocation and utilization of
limited-capacity attentional resources. A sufficient quantity of attentional resources must be available and allocated throughout the
duration of a task to reduce the intensity and adverse impact of negative thoughts and emotions. It is reasonable to expect protracted
and difficult performance tasks to consume a greater portion of these limited resources, particularly tasks involving threatening stimuli,
such as abuse froma higher-level manager. These situations are likely to evoke strong negative emotions of varying intensity and
frequency throughout the duration of a task, while continuing to drain limited attentional resource capacity. Hence, transformational
leadership relies on sufficient attentional capacity to maintain effective self-regulation, which mitigates the adverse impact of negative
emotions over the course of the task.We summarize this in the following hypothesis.
Hypothesis 1. Higher levels of attentional resource capacity will reduce the effect of pre-task negative emotions during demanding
performance tasks, resulting in effective self-regulation and leading to lower post-task negative emotions and higher levels of proactive
behavior and transformational leadership, or lower levels of abusive supervision.
If attentional resource capacity is insufficient to resolve the coactivation of dominant and non-dominant networks, for example,
those processing task-related and threat-related stimuli, then self-regulation will be ineffective, leading to an increase in negative
emotions. Any subsequent increase in negative emotions will further deplete limited attentional capacity, leading to weaker selfregulation
and abusive supervision, which we summarize in the following hypothesis.
Hypothesis 2. Lower levels of attentional resource capacity will intensify the effect of pre-task negative emotions during demanding
performance tasks, resulting in ineffective self-regulation and leading to higher post-task negative emotions and higher levels of
abusive supervision, or lower levels of proactive behavior and transformational leadership.
As stated previously, we do not make the same predictions for easy or simple performance tasks. Such tasks are unlikely to place
heavy demands on a leader’s cognitive and attentional resources in the same way as difficult or demanding tasks (Shamosh & Gray,
2007; Shamosh et al., 2008); hence, a leader’s attentional resource capacity is unlikely to diminish greatly over timewhen performing
a simple or easy task. Furthermore, easy tasks are not likely to be perceived as threatening and therefore threat-sensitive leaders will
have sufficient attentional resource capacity to process task-related thoughts. Indeed, negative emotions (e.g., anxiety) may interact
with motivational factors, such as extra effort, leading to improved performance on easy tasks (Eysenck et al., 2007).
In summary, we have argued that our process model of self-regulation and leadership predicts two separate and distinct forms of
leadership during demanding performance tasks, namely, transformational leadership and abusive supervision, which are forms of
constructive leadership and destructive leadership, respectively. Having developed our hypotheses, we next describe a “realworld”
field study that we conducted to test our model.
The participants in this study occupied formal leadership roles in several Australian organizations andwere attending a leadership
development program conducted by their organization. At the time of this study, opportunities to attend such programs were relatively
limited and selectionwas based largely onmerit. Hence,most participants were highlymotivated and effective in their current
role and were attending the program largely for promotional rather than remedial reasons. The first author was engaged by each organization
to conduct the program and provide participants with a face-to-face debrief of their test results. Ethics approval for the
study was provided by our University.
Drawing on prior field research experience, we anticipated a number of factors that might unexpectedly reduce the size of our
sample, for example, shifting organizational priorities or cuts to training budgets. Hence, we sought involvement from several organizations
to mitigate this risk. Importantly, our hypotheses required an experimental design, yet this study was to be conducted in
work situations where we had far less control than in the laboratory (Antonakis, Bendahan, Jacquart, & Lalive, 2010).We therefore
decided to randomly allocate participants to control versus experiment groups based on an approximate 1 to 3 ratio to maximize
our sample size for the experiment condition.
Testing our processmodel of self-regulation and leadership required a study design that could: (1)measure state-like constructs,
namely, attentional resource capacity, negative emotions and self-regulation; (2) induce self-regulatory impairment; and (3) quickly
gain acceptance at face-value from busy leaders and human resource managers who were indifferent to our research objectives.We
therefore designed a short and simple experimental procedure, utilizing familiar paper-and-pencil measures, which could be easily
incorporated in the leadership development programs conducted by the first author. Given the central importance and conceptualization
of self-regulation as a state-like construct,we next outline the procedurewe used tomanipulate and measure this, followed by
a description of our experimental procedure.
Most ego depletion studies use a number of common procedures for depleting self-control resources (e.g., Bertrams et al., 2013);
however, in these studies self-control exertion is usually measured by self-report questionnaires (see review by Hagger et al., 2010).
Some example items include: “How difficult did you find the task?” and “How exhausted do you feel right now?” (Bertrams et al.,
2013, p. 671). If, as Bertrams et al. (2013) suggest, “altering one’s behavior to comply with rules and standards is the essence of selfregulation”
(p. 670, emphasis added), then these self-report questionnaires do not actually measure changes in behavior, but simply
perception of task difficulty. In this studywemeasured self-regulation using a time-limited difficult mathematical test,which provided
a simple but effective way to deplete attentional resources and therefore impair self-regulation.
We reasoned that most participants in formal leadership roles have reasonably high personal performance standards or goals, as
these are essential in positions of leadership responsibility (Latham & Locke, 1991; Yeow & Martin, 2013). Accordingly, we reasoned
that time-limited mathematical tests involve “tradeoff” decisions between speed and accuracy to attain a test score congruent with
high personal performance expectations (seeWood & Jennings, 1976). Trade-off decisions are reflected in a participant’s behavioral
response to questions contained in the test. Specifically, personal standards or goals emphasizing accuracy over speed are observed in
a behavioral response where the ratio of errors committed to questions answered is relatively low. In contrast, personal standards or
goals emphasizing speed over accuracy are observed in a behavioral response where the ratio of errors committed to questions answered
is relatively high.We assign the term error rate to this behavioral tendency and define this as the ratio of errors committed
to the number of questions answered.
We further reasoned that if a participant found a mathematical test to be relatively easy then we would expect a behavioral response
that is rapid with a high number of accurate answers and relatively few errors (e.g., Dickman, 1990). In a time-limited test
this would equate to a low error rate (i.e., a low ratio of errors committed to questions answered). Also, it would be unlikely for participants
to feel overly threatened or experience increasing negative emotions over the course of an easy test. However, in a test perceived
as difficult,wewould expect a participant to adopt a slower,more deliberate and cautious behavioral response to achieve a test
score commensurate with high personal standards or goals. Again, in a time-limited test this should equate to a relatively low error
rate. This would reflect effective self-regulation.
However, in the same difficult test condition, if a participant responded rapidly and impulsively, committing a relatively higher
ratio of errors to questions answered, then this higher error rate would reflect ineffective self-regulation (as this behavioral response
entailsmakingmistakes). In addition, it is highly likely that this situationwould arouse negative emotions (e.g., anxiety, anger or fear)
thatwould increase over the course of the difficult test, further draining attentional resource capacity. Evidence supporting this effect
can be found in the literature linking cognitive resource depletion and negative emotions with mathematical errors (see Ashcraft &
Kirk, 2001; Ayres, 2001; Ayres & Sweller, 1990). Mathematical errors can occur through lack of knowledge, however, they can also
result from test anxiety and limitations in working memory, which is a form of attentional resource (Campbell, 1987; DeStefano &
LeFevre, 2004; Fayol, Abdi, & Gombert, 1987). Hence, we used a time-limited difficult mathematical test to initiate an “emotion driven
self-regulation failure” (MacCoon et al., 2004), while a participant’s level of self-regulation was measured via the error rate obtained
on the test.
Participants and study design
We collected data over a 15-month period from 187 organizational leaders who were attending a leadership development program
conducted by their organization.We invited their direct superior to rate them on an example of constructive and destructive
leadership, to minimize the potential effect of common-method variance (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). However,
26 (13.9%) superiors failed to rate their subordinate within a four-week timeframe and we therefore excluded those participants
without superior ratings from the study. Hence our final sample comprised 161 participants with matching superior ratings of leadership.
All participantswere full-timeleaders drawn fromsix organizations representing themining (61.5%), government (33.5%) and
construction (5%) sectors. Our sample comprised front-line supervisors (56.5%), managers of front-line supervisors (13.6%), business
unitmanagers (19.3%) and senior executives (10.6%). Themajorityweremale (85.6%), held a bachelor’s degree or higher (53.4%) and
were an average of 42.5 years old (standard deviation = 9.5 years).
On the first day of each leadership program attendees provided their consent to participate in the study and, unbeknownst to
them,were randomly allocated to a control group (easy mathematical test) or an experiment group (difficult mathematical test). Participants
then completed a number of pen-and-paper measures in the following order: (1) trait emotions; (2) attentional resource
capacity; (3) negative state emotions; (4) an easy or difficult mathematical test for the control or experiment condition, respectively;
and (5) negative state emotions (repeat-measure). Third-party ratings of constructive and destructive leadershipwere completed online
by each participant’s superior within four weeks of this procedure.
Constructs and measures
Attentional resource capacity
Ameasure of “fluid intelligence”was used as an index of attentional resource capacity. Fluid intelligence is defined as the ability to
reason and solve novel problems and has been associated with working memory, abstract reasoning and attention (Kanfer &
Ackerman, 2004). There is also evidence linking fluid intelligence to greater utilization of cognitive resources on difficult decisionmaking
tasks (Del Missier, Mantyla, & De Bruin, 2012; Shamosh & Gray, 2007). The Australian Council for Educational Research
(ACER) abstract reasoning test is a commonly used measure of fluid intelligence for personnel selection purposes. It has an internal
reliability of 0.83 (Morgan, Stephanou, & Simpson, 2000) andwas used to establish a baselinemeasure of attentional resource capacity.
Participants completed Part 1 of the test, which contains 20 multiple-choice questions assessing pattern recognition of abstract
figures. A high fluid intelligence score indicates high attentional resource capacity.
Negative state emotion
The Positive and Negative Affect Scale (PANAS) provides a measure of pleasurable and aversive emotional states, respectively
(Watson, Clark, & Tellegen, 1988). We used the 10-item Negative Affect (NA) scale (e.g., “distressed,” “upset,” “scared”) to measure
negative state emotion. Participantswere asked to rate on a five-point scale (1=very slightly or not at all, 5=verymuch) the extent
to which they experienced each particular emotion immediately before (pre-test) and immediately after completing (post-test) the
mathematical test.Watson et al. (1988) report an internal reliability of 0.85 for the NA scale. The PANAS has been used previously in
destructive leadership studies (e.g., Mawritz et al., 2014). High scores indicate high positive and negative emotionality.
Participants in the control condition (easy mathematical test) completed the ACER Test of Employment Entry Maths (TEEM),
which is a multiple-choice test of basic mathematical ability containing 32 questions. The time limit for the test is 25 min and it
has good internal reliability of 0.80 (Izard, Woff, & Doig, 1992). Participants in the experiment condition (difficult mathematical
test) completed the ACER Select Professional Numerical test (SPN), which is an advanced test of mathematical ability containing 29
questions that include number sequences, arithmetical reasoning and number matrices. The time limit for the test is 20 min and it
has good internal reliability of 0.80 (ACER, 2003). Self-regulatory impairmentwasmeasured by calculating the percentage of total errors
committed divided by the total number of questions answered by participants on the test (labeled error rate). As previously argued,
a low error rate reflects effective self-regulation and a high error rate reflects ineffective self-regulation.
Proactivity, transformational leadership and abusive supervision
The direct superiors of each participant were asked to rate the behavior of their subordinate over the previous month. The onemonth
timeframe was used to ensure an adequate sampling of recent participant behaviors displayed under varying contexts. Each
superior completed the nine-item proactivity scale of the Multidimensional Performance Questionnaire (MPQ: Griffin et al., 2007).
This scale consists of three sub-scales measuring proactive behavior at the individual, team and organizational levels. The measure
uses a five-point rating scale (1 = very little, 5 = a great deal) and has good internal reliability, ranging from 0.88 to 0.94 across
the three sub-scales. Example items include: “made changes to the way their core tasks are done,” “developed new and improved
methods to help their work unit perform better,” and “became involved in changes that are helping to improve the overall effectiveness
of the organization.” High scores indicate high proactive behavior.
Each superior also completed the 15-item measure of transformational leadership developed by Rafferty and Griffin (2004). This
contemporary measure of constructive leadership consists of five sub-factors of transformational leadership: vision, intellectual stimulation,
inspirational communication, supportive leadership and personal recognition. The measure uses a five-point rating scale
(1= strongly disagree, 5= strongly agree) and has good internal reliability, ranging from0.82 to 0.96 across the five sub-factors. Example
items include: “has a clear understanding ofwherewe are going,” “challenges others to think about old problems in newways,”
and “says positive things about the work unit.” High scores indicate high transformational leadership.
Finally, the same superior completed the 15-item abusive supervision questionnaire (Tepper, 2000), which is a well-established
measure of destructive leadership (Hershcovis & Rafferty, 2012). Themeasure uses a five-point rating scale (1=I cannot remember
him/her ever using this behavior, 5 = he/she uses this behavior very often) and has good internal reliability of 0.90. Example items
include: “ridicules others,” “tells others their thoughts or feelings are stupid,” and “gives others the silent treatment.” High scores indicate
high abusive supervision.
Negative trait emotions
We measured participants’ negative trait emotions before starting the procedure to compare the general tendency to experience
negative state emotions between participants in the control and experiment conditions.We reasoned that any significant difference
between the groupsmight influence our results. The trait anxiety, anger and immoderation (i.e., impulsivity) scales fromthe International
Personality Item Pool (IPIP; Goldberg, 1990, 1992) were used for this purpose. Each scale contains 10 items and participants
were asked to rate howaccurately each itemdescribed themin general (1=very inaccurate, 5=very accurate). The IPIP is a widely
used public domain personality measure and scales have good internal reliabilities of 0.83, 0.88 and 0.77, for trait anxiety, anger and
immoderation, respectively (International Personality Item Pool). High scores indicate high negative trait emotions.
We adopted a strictly confirmatory framework where we proposed a single theory-derived model and tested the fit of thismodel
to the control data (easy mathematical test) and experiment data (difficult mathematical test) as recommended by Joreskog (1993).
Our hypotheseswere tested using path analysis (see Fig. 1) and hierarchical regression analysis to model the simple slopes of the proposed
Table 1 shows the means (M), standard deviations (SD), reliability coefficients (α) and analysis of variance between variables in
the control and experiment conditions. Zero-order correlations for all variables are presented separately for each condition in
Table 2. The reliability coefficients are generally acceptable (between 0.66 and 0.96) and the correlations between key variables support
the expected relationships described in our model, which we summarize next.
Examining Table 2, in the control condition (n = 38), attentional resource capacity is negatively correlated with self-regulatory
impairment (r=−.66, p b .01) and positively correlatedwith superior-rated proactivity (r=.33, p b .05); pre-test negative emotion
has a weak negative correlation with self-regulatory impairment (r = −.41, p b .10); pre- and post-test negative emotion are positively
correlated (r = .33, p b .05); and self-regulatory impairment has a weak positive correlation with post-test negative emotion
(r=.31, p b .10) and is negatively correlatedwith proactivity (r=−.35, p b .05).Hence, participantswith higher attentional resource
capacity demonstrate effective self-regulation during an easy mathematical test and are rated higher in proactivity by their superior.
Furthermore, participantswith higher levels of pre-test negative emotion are likely to demonstrate effective self-regulation (although
this effect is weak). Higher levels of pre-test negative emotion and effective self-regulation are also associated with higher scores on
the easymathematical test (r=.40, p b .05 and r=−.94, p b .01, respectively). Finally, effective self-regulation during the easy mathematical
test is associated with higher superior-rated proactivity, which is strongly correlated with higher transformational leadership
(r = .62, p b .001).
In the experiment condition (n = 123), attentional resource capacity is negatively correlated with self-regulatory impairment
(r=−.41, p b .01) and negatively correlated with superior-rated abusive supervision (r=−.19, p b .05); pre-test negative emotion
is positively correlated with self-regulatory impairment (r = .21, p b .01); pre- and post-test negative emotion are positively correlated
(r = .47, p b .01); and self-regulatory impairment is positively correlated with post-test negative emotion (r = .37, p b .01)
and with abusive supervision, although weakly (r = .15, p b .10), and negatively correlated with proactivity (r = −.20, p b .05).
Hence, participants with higher attentional resource capacity also demonstrate effective self-regulation during a difficult mathematical
test and are rated lower in abusive supervision by their superior. However, participants with higher levels of pre-test negative
emotion are likely to demonstrate ineffective self-regulation, which is associated with lower scores on the difficult mathematical
test (r = −.73, p b .01). Consistent with participants in the control condition, an increase in the level of pre-test negative emotion
leads to a significant increase in post-test negative emotion via self-regulatory impairment. Furthermore, ineffective self-regulation
during the difficult mathematical test predicts higher superior-rated abusive supervision, although this effect is weak, and lower
superior-rated proactivity (r=−.20, p b .05). Finally, superior-rated proactivity is strongly correlated with higher transformational
leadership (r = .66, p b .001).
Finally, the relationship between negative trait emotions and superior-rated leadership is generally consistent with expectations
(particularly in the larger experiment group). For example, trait anxiety, anger and impulsivity are negatively correlated with transformational
leadership, and positively correlated with abusive supervision. Also, the relationship between constructive and destructive
forms of leadership is consistent with expectations, i.e., transformational leadership and abusive supervision are negatively
correlated (r=−.31, p b .10 in the control condition, and r=−.50, p b .001 in the experiment condition). Lastly, there is no significant
correlation between post-test negative emotion and superior-rated proactivity, transformational leadership or abusive supervision
in either condition.
We first compared participant performance on the mathematical test between the control and experiment conditions. The ANOVA
results in Table 1 indicate that performance on the test differed significantly between the two conditions. In contrast to the experiment
condition, participants in the control condition: (1) answeredmore questions (M= 96.79, SD=6.48versusM= 54.33, SD= 17.32; F(1,
159)= 216.14, p b .001); (2) attained a higher test score (M= 86.10, SD= 12.64 versus M= 35.80, SD= 17.11; F(1, 159)= 277.87,
p b .001); and (3) achieved a lower error rate (M= 11.39, SD=9.20 versusM= 35.00, SD= 20.89; F(1, 159)= 45.10, p b .001). Furthermore,
participants in the experiment condition reported a significantly higher level of negative emotion following the test than those
in the control condition (M= 1.57, SD = .51 versus M= 1.21, SD=.23; F(1, 159)= 17.39, p b .001). Finally, there was no significant
difference in attentional resource capacity between participants in the control and experiment conditions (M= 13.63, SD=3.31 versus
M = 12.59, SD = 3.78; F(1, 159) = 2.47, p= .118).
Next,we examined the difference in negative emotions pre- and post-test in each condition. The level of negative emotion for participants
in the control condition reduced slightly over the duration of themathematical test (fromM= 1.29, SD=.27 toM= 1.21,
SD= .23), however, this reduction was non-significant (t(37)= 1.60, p = .118). Conversely, participants in the experiment condition
experienced a large and significant increase in the level of negative emotion (from M= 1.17, SD = .24 to M= 1.57, SD = .51;
t(122)=−9.70, p b .001). Finally, it is unlikely that any difference in state emotion can be attributed to differences in trait emotions
between participants in each condition. Table 2 shows that the levels of trait anxiety, anger and impulsivity are similar between the
Together, these results indicate that participants have similar levels of attentional resource capacity and negative trait emotions,
yet those in the experiment condition performedworse on the difficult mathematical test and experienced higher negative emotions
post-test than participants in the control condition, who completed the easier mathematical test. Furthermore, the slight decrease in
negative emotions over the duration of the easiermathematical test (control condition) suggests that participants found this test to be
relatively easy and therefore non-threatening. However, it appears that participants in the experiment condition found their mathematical
test to be difficult and more threatening, leading to a significant increase in negative emotion over the duration of the test.
These results provide initial support for our argument that leaders who undertake a demanding performance task experience an
increase in negative emotions over the duration of that task. Specifically, given that participants in both conditions had equivalent
attentional resource capacity immediately before the test, it is plausible that the more cognitively demanding test depleted this
resource over time, thus accounting for higher self-regulatory impairment and negative emotions in the experiment group only.
We explore this further in the following path analysis.
We report the standardized regression weights for each path and the squared multiple correlations for the variables in ourmodel
for both the control and experiment conditions in the prediction of transformational leadership and abusive supervision (see Figs. 2
and 3). We next report the following goodness-of-fit indexes for each path model and condition. First, the hypothesized superiorrated
proactivity–transformational leadership path model (Fig. 2) was an acceptable fit to the data in both the control, χ2(10, N =
38) = 13.71, p = .19; CFI = .94; RMSEA = .10, and experiment conditions χ2(10, N = 123) = 16.87, p = .08; CFI = .99;
RMSEA = .07. Second, the hypothesized superior-rated abusive supervision path model (Fig. 3) was a good fit to the data in the
control condition, χ2(10, N = 38) = 12.11, p = .28; CFI = .99; RMSEA = .03, and a satisfactory fit in the experiment condition
χ2(10, N=123)= 16.56, p= .09; CFI =.97; RMSEA =.05.We have adopted the goodness-of-fit indexes and cutoff values recommended
by Byrne (2010), namely, the chi-square (χ2) statistic (p N .05), comparative fit index (CFI) N .95, and rootmean square error
of approximation (RMSEA) b .05.While the RMSEA exceeds the recommended cut-off value for the superior-rated proactivity–transformational
leadership pathmodel (Fig. 2), on balance we consider the goodness-of-fit indexes (particularly the chi-square statistic)
to be acceptable for this model.
To testwhether specific parameters in our two pathmodelsdifferedsignificantly between the control and experiment conditions,
we next conducted a multi-group comparison in which all regression weights and covariances in each model were constrained to be
equal for both conditions (as recommended in Byrne, 2010). In addition, all predictors were mean-centered prior to analysis to minimize
the potential effects ofmulticollinearity (see Little, Card, Bovaird, Preacher, & Crandall, 2007). The constrained model fit statistic
for the proactivity–transformational leadership path model (Fig. 2) was significantly different, χ2(31, N = 161) = 60.72, p = .001,
fromthe unconstrained model, χ2(20, N=161)= 30.70, p=.06; CFI=.95; RMSEA=.06.While the constrained model fit statistic
for the superior-rated abusive supervision path model (Fig. 3) was also significantly different, χ2(20, N = 161) = 40.93, p = .004,
fromthe unconstrainedmodel, χ2(10, N=161)= 12.95, p=.23; CFI=.98; RMSEA=.04. These results indicate that the parameters
in each model differed significantly between the control and experiment conditions, suggesting that different effectswere obtained in
each condition. We next examine these direct and indirect effects in further detail.
Fig. 2 shows a significant direct effect of self-regulatory impairment on superior-rated proactivity in both the experiment (β =
−.24, p b .01) and control (β = −.34, p b .05) conditions. We interpret this as lower self-regulatory impairment, i.e., effective selfregulation
influences higher superior-ratings of proactivity, regardless of test difficulty. The indirect effects for this path model are
displayed in Table 3.
There are significant indirect effects through self-regulation impairment from pre-test negative emotion to superior-rated
proactivity and transformational leadership in both the control and experiment conditions (β = .06, p b .05 and β = −.03, p b .05,
respectively, as shown in Table 3). Examining Fig. 2 reveals that higher pre-test negative emotion influences higher superior-rated
proactivity and transformational leadership, through effective self-regulation (i.e., lower error rate) in the easy test condition
(i.e., lower cognitive load). Conversely, in the difficult test condition (i.e., higher cognitive load), higher pre-test negative emotion
influences lower superior-rated proactivity and transformational leadership, through ineffective self-regulation, as predicted
(i.e., higher error rate).
Second, there are significant indirect effects through self-regulation impairment from attentional resource capacity to superiorrated
proactivity and transformational leadership in both the control and experiment conditions (β = .12, p b .05 and β = .07,
p b .01, respectively, as shown in Table 3). That is, higher attentional resource capacity influences effective self-regulation
(i.e., lower error rate) and higher superior-rated proactivity and transformational leadership, regardless of test difficulty, as predicted.
Third, there are significant indirect effects through self-regulation impairment fromattentional resource capacity to post-test negative
emotion in the control and experiment conditions (β=−.31, p b .01 and β=−.12, p b .01), respectively, as shown in Table 3.
That is, higher attentional resource capacity, through effective self-regulation (i.e., lower error rate), influences lower post-test negative
emotion, as predicted.
Fourth, there are significant indirect effects through self-regulation impairment between pre- and post-test negative emotions in
the control and experiment conditions (β=−.16, p b .01 and β = .05, p b .05, respectively, as shown in Table 3). Fig. 2 indicates that
effective self-regulation (i.e., lower error rate) leads to a reduction in negative emotions over the duration of the easy test, whereas
ineffective self-regulation (i.e., higher error rate) influences an increase in negative emotions over the course of the difficult test, as
Finally, there is a significant indirect effect through self-regulation impairment fromthe two-way interaction between attentional
resource capacity and pre-test negative emotion, to superior-rated proactivity and transformational leadership in the experiment condition
only (β=.15, p b .05, as shown in Table 3). It is difficult to interpret the indirect effect of this two-way interaction on superiorrated
transformational leadership. For example, the indirect effect of this two-way interaction on superior-rated proactivity is nonsignificant
(β = −.02, p = .25), while a significant indirect effect exists through superior-rated proactivity from self-regulation
impairment to superior-rated transformational leadership in both the control and experiment conditions (β = −.22, p b .05 and
β = −.16, p b .01, respectively, as shown in Table 3). We interpret this to mean that effective self-regulation leads to higher
superior-rated transformational leadership, through superior-rated proactivity, regardless of the level of test difficulty.
Fig. 3 indicates a significant direct effect of self-regulatory impairment on superior-rated abusive supervision in the experiment
condition (β = .19, p b .05), but not the control condition (β=.15, p= .37).We interpret this as ineffective self-regulation leading
to higher superior-rated abusive supervision, but only in the difficult test condition (i.e., high cognitive load), as predicted.We next
examine the indirect effects for this path model (see Table 3).
First, there is a significant indirect effect through self-regulation impairment from pre-test negative emotion to superior-rated
abusive supervision for the experiment condition (β = .04, p b .05), but not the control condition (β = −.04, p = .45), as shown
in Table 3. Examining Fig. 3 reveals that higher pre-test negative emotion leads to higher superior-rated abusive supervision, via ineffective
self-regulation (i.e., higher error rate) in the difficult test condition only (i.e., high cognitive load).
Second, Table 3 indicates that a significant indirect effect exists through self-regulation impairment from attentional resource capacity
to superior-rated abusive supervision for the experiment condition (β = −.08, p b .05), but not the control condition (β =
−.09, p = .44). This indicates that higher levels of attentional resource capacity lead to effective self-regulation (i.e., lower error
rate) and lower superior-rated abusive supervision in the difficult test condition only, as predicted. Finally, the indirect effects through
self-regulation impairment fromattentional resource capacity to post-test negative emotion, and between pre- and post-test negative
emotions, are the same as the superior-rated proactivity–transformational path model (Fig. 2).
Hierarchical regression analysis to interpret the simple slopes
Fig. 2 shows a significant direct effect of the two-way interaction between attentional resource capacity and pre-test negative emotion
on superior-rated proactivity for the experiment condition (β=.26, p b .01), but not for the control condition (β=−.02, p = .93). To
examine the effect of this interaction (in the difficult test condition only) we conducted a hierarchical regression analysis in which
proactivity was predicted by the main effects of attentional resource capacity and pre-test negative emotion at Step 1. We added a
two-way interaction between the mean-centered scores of attentional resource capacity and pre-test negative emotion at Step 2, as recommended
by Aiken andWest (1999). The results of this hierarchical regression analysis explained a significant increase in the variance
in superior-rated proactivity (ΔR2=.07, F(1, 119)= 3.38, p b .05). Since there are no significant indirect effects through self-regulation
fromthis two-way interaction to proactivity, as reported previously, we conducted simple slope analyses to examine the direct effect of
this interaction (see Little et al., 2007). Fig. 4a shows that, in linewith our hypothesis, high pre-test negative emotion significantly reduces
superior-rated proactivity among participantswho are low in attentional resource capacity (one SD below the mean), β=−1.53, p b .01.
However, superior-rated proactivity is higher among participants with high pre-test negative emotion and high attentional resource
capacity; however, this effect is not significant, β=−.43, p =.28.
Fig. 3 shows a significant direct effect of the two-way interaction between attentional resource capacity and pre-test negative emotion
on superior-rated abusive supervision for the experiment condition (β=−.29, p b .01), but not for the control condition (β=
−.16, p= .34). To examine the effect of this interaction (in the difficult test condition only) the same hierarchical regression analysis
as before was conducted, except that superior-rated abusive supervision (not proactivity) was predicted by themain effects of attentional
resource capacity and pre-test negative emotion at Step 1. The results of this hierarchical regression analysis explained a significant
increase in the variance in superior-rated abusive supervision (ΔR2 = .09, F(1, 119) = 5.72, p b .001). Again, there are no
significant indirect effects between this two-way interaction and abusive supervision (via self-regulation), so we conducted simple
slope analyses to examine the direct effect of this interaction. Fig. 4b shows that, in line with our hypothesis, high pre-test negative
emotion significantly increases superior-rated abusive supervision among participants who are low in attentional resource capacity
(one SD below the mean), β = .88, p b .001. However, superior-rated abusive supervision is lower among participants with high
pre-test negative emotion and high attentional resource capacity; however, this effect is not significant, β = .29, p = .15.
Our study provides evidence favoring a process model of leadership. Specifically, we demonstrate how tasks of varying difficulty
impact a leader’s attentional resource capacity and negative emotions, leading to differences in self-regulation and constructive or destructive
forms of leader behavior. In support of Hypothesis 1,we found that higher levels of attentional resource capacity reduced the
adverse impact of pre-task negative emotions during both difficult and easy performance tasks. This resulted in effective selfregulation,
leading to lower post-task negative emotion and higher levels of superior-rated proactivity and transformational leadership
(a form of constructive leadership). Although we did not formally postulate the influence of easy performance tasks on selfregulation,
our results demonstrate the impact of varying contextual factors on the relationship between pre-task negative emotions
and self-regulation. We found that high pre-task negative emotion in the difficult test condition led to ineffective self-regulation,
whereas high pre-task negative emotion in the easy test condition led to effective self-regulation. Furthermore, our results suggest
that self-regulation influences transformational leadership indirectly through proactivity.We believe that these results might explain
the previously raised inconsistencies in studies examining the relationship between negative trait emotions and transformational
leadership (e.g., Bono & Judge, 2004; Judge & Bono, 2000).
Our results provide evidence that contextual factors such as task difficulty can influence the relationship between negative state
emotions, self-regulation and leader behavior (e.g., Eysenck et al., 2007). A more cognitively demanding task (i.e., the difficult test)
increased the adverse effect of negative emotions on attentional resource capacity, leading to ineffective self-regulation, lower
proactivity and lower transformational leadership. In contrast, a less cognitively demanding task (i.e., the easy test) decreased the adverse
effect of negative state emotion on attentional resource capacity, leading to effective self-regulation, higher proactivity and
higher transformational leadership. The majority of leadership trait studies do not examine the interaction of contextual factors,
such as task difficulty, and state-like leader attributes such as self-regulation and proactivity (see Avolio, 2007; Zaccaro, 2007,
2012). Therefore, the relationship between negative trait emotions and transformational leadership is likely to be confounded in studies
where these factors are ignored (Tett & Guterman, 2000).
Furthermore, it has been suggested that transformational leadership is more prevalent in dynamic rather than stable work environments
(Bass, 1990b; De Hoogh et al., 2005). However,we found a significant indirect effect, via proactivity, fromself-regulation to
transformational leadership in both the difficult and easy test conditions. That is, effective self-regulation predicted higher proactivity,
whichwas associatedwith higher transformational leadership regardless of the context.We do not suggest that the difficult and easy
mathematical tests used in our study represent dynamic and stable work environments, respectively. However, it is plausible that
theymight exert a similar influence on a leader’s psychological resources. For example, dynamicwork environments can be cognitively
and emotionally demanding (Finkelstein & Hambrick, 1996) in a similar way to difficult mathematical tests (Ayres, 2001). Hence,
our results suggest that transformational leadership might occur equally in dynamic and stable work environments, while effective
self-regulation and proactivity appear to be important antecedents underrepresented in the literature.
In support of Hypothesis 2, we found that lower levels of attentional resource capacity increased the adverse impact of pre-task
negative emotions during the difficult performance task only. This resulted in ineffective self-regulation, leading to higher posttask
negative emotions and higher levels of abusive supervision (a formof destructive leadership). Furthermore,we found that higher
levels of attentional resource capacity resulted in effective self-regulation and lower abusive supervision.
Common to both hypotheses, our results demonstrated that effective self-regulation relied on sufficient attentional resource capacity
regardless of the level of task difficulty. Higher attentional resource capacity resulted in effective self-regulation and lower
post-task negative emotion in both test conditions. Also, effective self-regulation resulted in reduced negative emotions, while ineffective
self-regulation resulted in increased negative emotions over the course of an easy or difficult performance task.
Finally, in terms of self-regulation, the amount of variance explained in the difficult test condition (R2= .227, p b .01) was less than
half that of the easy test condition (R2= .528, p b .01). This indicates the influence of other factors, in addition to attentional resource
capacity and pre-test negative emotion, during the difficult test condition. Perhaps the higher cognitive demands of the difficult test,
combined with increased negative emotions, depleted limited attentional resources and initiated increased effort or more deliberate
self-control strategies (e.g., Humphreys & Revelle, 1984).We leave this for future researchers to investigate.
Strengths, limitations and future research
We believe that our study makes a number of unique contributions to the leadership literature. First, it enhances our understanding
of why negative contextual factors like demanding and stressful performance tasks, common in dynamic work environments,
might result in both constructive and destructive leader behaviors. Our process model of self-regulation and leadership highlights
the central role of attentional resource capacity in reducing the adverse impact of negative emotions on self-regulation in these situations.
We have tested thismodel in a real-world setting with motivated leaders and demonstrated how this can predict both transformational
leadership and abusive supervision as examples of constructive leadership and destructive leadership, respectively.
Second, we have conceptualized self-regulation as a dynamic state-like construct that describes the capacity for leaders to exercise
self-control. In this way we have moved beyond static trait-like models to provide a more realistic account of how leader emotions
and behavior evolve over the course of an easy or difficult performance task.
Our study also has potential applicability beyond leadership research.We have developed and tested a simple and practicalway to
manipulate and measure self-regulation that is impervious to the influence of “response distortion,” i.e., attenuation of undesirable
traits like anger, and inflation of desirable traits like conscientiousness (see Bolino & Turnley, 2003; Tett & Simonet, 2011;
Vasilopoulos, Reilly, & Leaman, 2000). Response distortion can not only confound the interpretation of empirical results (e.g., De
Hoogh et al., 2005; Dobson, 2000), but weaken confidence in leadership selection decisions (see Hogan, Hogan, & Roberts, 1996;
Spillane, 2012).We believe that our experimental procedure could be used in organizations to identify those leaders who are more
vulnerable to self-regulatory impairment and destructive leadership, thereby reducing the prevalence and negative impact of such
harmful behavior. This could be achieved bymatching leaders to roles best suited to their capacity for self-control, and by specifically
focusing on emotional self-regulation training (Yeow & Martin, 2013).
The findings of this study should also be considered in light of several limitations, each of which should be addressed by future
research. First, the sample size of the easy test condition was considerably smaller than that of the difficult test condition (n = 38
and n=123, respectively).We made this decision due to our lack of control in a real-world experimental field study (versus laboratory
study) and attempted to maximize the size of the sample in the difficult test condition as thiswas our focal interest. Similar sized
samples have also been reported in recent leadership studies employing path analysis or structural equationmodeling (e.g., Hur, van
den Berg, &Wilderom, 2011; Nohe, Michaelis, Menges, Zhang, & Sonntag, 2013).
Second, we cannot say for certain whether a test of fluid intelligence provides the most appropriate measure of attentional resource
capacity. Working memory tests have also been linked to regions of the medial prefrontal cortex (Bush, Luu, & Posner,
2000; Fan et al., 2009; Rothbart, Sheese, & Posner, 2007) and have been associated with abstract reasoning and attention (Kanfer &
Ackerman, 2004). Future researchmight attempt to replicate our study using a suitable test ofworkingmemory tomeasure attentional
resource capacity (e.g., Ackerman, Beier, & Boyle, 2005; Hamilton, Hockey, & Rejman, 1977).
Finally,while ourmeasure of self-regulationmight be impervious to self-response inflation, self-reportmeasures of negative emotion
are not and can result in non-normal data that might confound the interpretation of results (e.g., Crawford & Henry, 2004;
Dobson, 2000). Although not a major issue in this study, this does highlight a potential risk in future research using self-report emotion
While leaders have the potential to inspire their subordinates and achieve ambitious organizational and social goals, the evidence
suggests that they can also cause unintentional harm and irreparable damage. In this paper we have attempted to explain one of the
reasons for this anomaly.We have presented and tested amodel of self-regulation and leadership that describes the process bywhich
both constructive and destructive leadership occur, and the context in which leaders are most vulnerable to destructive forms of behavior.
It is our hope that this study encourages further research into the antecedents of destructive leadership and effective remedies
to reduce its prevalence and adverse impact on individuals, organizations and society in general.
Aasland, M. S., Skogstad, A., Notelaers, G., Nielsen,M. B., & Einarsen, S. (2010). The prevalence of destructive leadership behaviour. British Journal of Management, 21(2),
ACER (2003). Acer select manual. ACER Press.
Ackerman, P. L., Beier, M. E., & Boyle, M. O. (2005). Working memory and intelligence: The same or different constructs? Psychological Bulletin, 131(1), 30–60.
Aiken, L. S., & West, S. G. (1999). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.
Anderson, C. A., & Bushman, B. J. (2002). Human aggression. Annual Review of Psychology, 53, 27–51.
Antonakis, J., Bendahan, S., Jacquart, P., & Lalive, R. (2010w). On making causal claims: A review and recommendations. Leadership Quarterly, 21(6), 1086–1120.
Ashcraft, M. H.,&Kirk,E. P. (2001). The relationships amongworkingmemory,math anxiety, and performance. Journal of Experimental Psychology-General, 130(2), 224–237.
Avolio, B. J. (2007). Promoting more integrative strategies for leadership theory-building. [Article]. American Psychologist, 62(1), 25–33.
Ayres, P. L. (2001). Systematic mathematical errors and cognitive load. Contemporary Educational Psychology, 26, 227–248.
Ayres, P. L., & Sweller, J. (1990). Locus of difficulty in multistage mathematics problems. The American Journal of Psychology, 103(2), 167–193.
Baskin-Sommers, A. R.,Wallace, J. F.,MacCoon, D. G., Curtin, J. J., & Newman, J. P. (2010). Clarifying the factors that undermine behavioral inhibition systemfunctioning
in psychopathy. Personality Disorders: Theory, Research, and Treatment, 1(4), 203–217.
Bass, B. M. (1990a). Bass & Stogdill’s handbook of leadership: Theory, research, and managerial applications (Third ed.). New York: Macmillan, Inc.
Bass, B. M. (1990b). From transactional to transformational leadership: Learning to share the vision. Organizational Dynamics, 18(3), 19–31.
Bass, B. M., & Avolio, B. J. (1997). Full range leadership development: Manual for the multifactor leadership questionnaire. Palo Alto, California: Mind Garden.
Baumeister, R. F., Bratslavsky, E., Muraven, M., & Tice, D. M. (1998). Ego depletion: Is the active self a limited resource? Journal of Personality and Social Psychology,
Baumeister, R. F., Vohs, K. D., & Tice, D. M. (2007). The strength model of self-control. Current Directions in Psychological Science, 16(6), 351–355.
Beal, D. J., Weiss, H. M., Barros, E., & MacDermid, S. M. (2005). An episodic process model of affective influences on performance. Journal of Applied Psychology, 90(6),
Bertrams, A., Englert, C., Dickhauser, O., & Baumeister, R. F. (2013). Role of self-control strength in the relation between anxiety and cognitive performance. Emotion,
Bolino, M. C., & Turnley,W. H. (2003). More than one way to make an impression: Exploring profiles of impression management. Journal of Management, 29(2), 141–160.
Bono, J. E., & Judge, T. A. (2004). Personality and transformational and transactional leadership: A meta-analysis. Journal of Applied Psychology, 89(5), 901–910.
Botvinick, M. M., Braver, T. S., Barch, D. M., Carter, C. S., & Cohen, J. D. (2001). Conflict monitoring and cognitive control. Psychological Review, 108(3), 624–652.
Bush, G., Luu, P., & Posner, M. I. (2000). Cognitive and emotional influences in anterior cingulate cortex. Trends in Cognitive Sciences, 4(6), 215–222.
Byrne, B. M. (2010). Structural equation modeling with AMOS: Basic concepts, applications, and programming (2nd ed.). New York, NY: Routledge/Taylor & Francis Group, US.
Campbell, J. I. D. (1987). Network interference and mental multiplication. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13(1), 109–123.
Cervone, D., Shadel, W. G., Smith, R. E., & Fiori, M. (2006). Self-regulation: Reminders and suggestions from personality science. Applied Psychology-An International
ReviewPsychologie Appliquee-Revue Internationale, 55(3), 333–385.
Costa, P. T., & McCrae, R. R. (1991). Revised neo personality inventory (NEO-PI-R) and NEO five-factor inventory (NEO-FFI): Professional manual. Odessa, FL: Psychological
Crawford, J. R., & Henry, J. D. (2004). PANAS construct validity. British Journal of Clinical Psychology, 43, 245–265.
Crossley, C. D., Cooper, C. D., & Wernsing, T. S. (2013). Making things happen through challenging goals: Leader proactivity, trust, and business-unit performance.
Journal of Applied Psychology, 98(3), 540–549.
De Hoogh, A. H. B., Den Hartog, D. N., & Koopman, P. L. (2005). Linking the big five-factors of personality to charismatic and transactional leadership; perceived dynamic
work environment as a moderator. Journal of Organizational Behavior, 26(7), 839–865.
Del Missier, F., Mantyla, T., & De Bruin,W. B. (2012). Decision-making competence, executive functioning, and general cognitive abilities. Journal of Behavioral Decision
Making, 25(4), 331–351.
Derryberry, D., & Reed, M. A. (2002). Anxiety-related attentional biases and their regulation by attentional control. Journal of Abnormal Psychology, 111, 225–236.
DeStefano, D., & LeFevre, J. A. (2004). The role of working memory in mental arithmetic. European Journal of Cognitive Psychology, 16(3), 353–386.
Dickman, S. J. (1990). Functional and dysfunctional impulsivity: Personality and cognitive correlates. Journal of Personality and Social Psychology, 58(1), 95–102.
Dinh, J. E., & Lord, R. G. (2012). Implications of dispositional and process views of traits for individual difference research in leadership. The Leadership Quarterly, 23(4),
Dobson, P. (2000). An investigation into the relationship between neuroticism, extraversion and cognitive test performance in selection. International Journal of
Selection and Assessment, 8(3), 99–109.
Einarsen, S., Aasland, M. S., & Skogstad, A. (2007). Destructive leadership behaviour: A definition and conceptual model. Leadership Quarterly, 18(3), 207–216.
Eubanks, D. L., & Mumford, M. D. (2010a). Leader errors and the influence on performance: An investigation of differing levels of impact. The Leadership Quarterly, 21,
Eubanks, D. L., & Mumford, M. D. (2010b). Destructive leadership: The role of cognitive processes. In B. Schyns, & T. Hansbrough (Eds.), When leadership goes wrong:
Destructive leadership, mistakes, and ethical failures (pp. 23–47).
Eysenck, M. W., & Calvo, M. G. (1992). Anxiety and performance: The processing efficiency theory. Cognition & Emotion, 6(6), 409–434.
Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive performance: Attentional control theory. Emotion, 7(2), 336–353.
Fan, J., Gu, X., Liu, X., Fossella, J., Wang, H., & Posner, M. I. (2009). Testing the behavioral interaction and integration of attentional networks. Brain and Cognition, 70,
Fayol, M., Abdi, H., & Gombert, J. E. (1987). Arithmetic problems formulation and working memory load. Cognition and Instruction, 4(3), 187–202.
Finkelstein, S., & Hambrick, D. C. (1996). Strategic leadership: Top executives and their effects on organizations. St Paul, MN: West.
Franke, F., & Felfe, J. (2011). How does transformational leadership impact employees’ psychological strain? Examining differentiated effects and the moderating role
of affective organizational commitment. Leadership, 7(3), 295–316.
Goldberg, L. R. (1990). An alternative “description of personality”: The big-five factor structure. Journal of Personality and Social Psychology, 59(6), 1216–1229.
Goldberg, L. R. (1992). The development of markers for the big-five factor structure. Psychological Assessment, 4(1), 26–42.
Griffin, M. A., Neal, A., & Parker, S. K. (2007). A new model of work role performance: Positive behavior in uncertain and interdependent contexts. Academy of
Management Journal, 50(2), 327–347.
Hagger, M. S., Wood, C., Stiff, C., & Chatzisarantis, N. L. D. (2010). Ego depletion and the strength model of self-control: A meta-analysis. Psychological Bulletin, 136(4),
Hambrick, D. C. (1989). Putting top managers back in the strategy picture. Strategic Management Journal, 10, 5–15.
Hambrick, D. C., Finkelstein, S., & Mooney, A. C. (2005). Executive job demands: New insights for explaining strategic decisions and leader behaviors. Academy of
Management Review, 30(3), 472–491.
Hamilton, P., Hockey, G. R. J., & Rejman, M. (1977). The place of the concept of activation in human information processing theory: An integrative approach. In S. Dornic
(Ed.), Attention and performance, VI, Hillsdale, N. J.: Erlbaum.
Hardy, L., Beattie, S., & Woodman, T. (2007). Anxiety-induced performance catastrophes: Investigating effort required as an asymmetry factor. British Journal of
Psychology, 98, 15–31.
Hershcovis, M. S., & Rafferty, A. E. (2012). Predicting abusive supervision. In J. Houdmont, S. Leka, & R. Sinclair (Eds.), (First ed.). Contemporary occupational health
psychology: Global perspectives on research and practice, Vol. 2. John Wiley & Sons, Ltd.
Hogan, R., Hogan, J., & Roberts, B. W. (1996). Personality measurement and employment decisions. American Psychologist, 51(5), 469–477.
Humphreys, M. S., & Revelle, W. (1984). Personality, motivation, and performance: A theory of the relationship between individual differences and information
processing. Psychological Review, 91(2), 153–184.
Hur, Y., van den Berg, P. T., & Wilderom, C. P. (2011). Transformational leadership as a mediator between emotional intelligence and team outcomes. The Leadership
Quarterly, 22(4), 591–603.
Izard, J., Woff, I., & Doig, B. (1992). Test of employment entry mathematics. Melbourne, Victoria: ACER.
International Personality Item Pool: A Scientific Collaboratory for the Development of Advanced Measures of Personality Traits and Other Individual Differences
(http://ipip.ori.org/). Internet Web Site.
Joosten, A., van Dijke, M., Van Hiel, A., & De Cremer, D. (2014). Being “in control” may make you lose control: The role of self-regulation in unethical leadership
behavior. Journal of Business Ethics, 121(1), 1–14.
Joreskog, K. G. (1993). Testing structural equation models. Testing structural equation models. Newbury Park, CA: Sage, 294–316.
Judge, T. A., & Bono, J. E. (2000). Five-factor model of personality and transformational leadership. Journal of Applied Psychology, 85(5), 751–765.
Kanfer, R., & Ackerman, P. L. (2004). Aging, adult development, and work motivation. Academy of Management Review, 29(3), 440–458.
Kant, L., Skogstad, A., Torsheim, T., & Einarsen, S. (2013). Beware the angry leader: Trait anger and trait anxiety as predictors of petty tyranny. The Leadership Quarterly,
Krasikova, D. V., Green, S. G., & LeBreton, J. M. (2013). Destructive leadership: A theoretical review, integration, and future research agenda. Journal of Management,
Latham, G. P., & Locke, E. A. (1991). Self-regulation through goal setting. Organizational Behavior and Human Decision Processes, 50(2), 212–247.
Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer.
Lim, B. C., & Ployhart, R. E. (2004). Transformational leadership: Relations to the five-factor model and team performance in typical and maximum contexts. Journal of
Applied Psychology, 89(4), 610–621.
Little, T. D., Card, N. A., Bovaird, J. A., Preacher, K. J., & Crandall, C. S. (2007). Structural equation modeling of mediation and moderation with contextual factors. Modeling
contextual effects in longitudinal studies, 207–230.
MacCoon, D. G.,Wallace, J. F., & Newman, J. P. (2004). Self-regulation: The context-appropriate allocation of attentional capacity to dominant and nondominant cues. In
R. F. Baumeister, & K. D. Vohs (Eds.), Handbook of self-regulation research (pp. 422–444). New York: Guilford Press.
Mawritz, M. B., Folger, R., & Latham, G. P. (2014). Supervisors’ exceedingly difficult goals and abusive supervision: Themediating effects of hindrance stress, anger, and
anxiety. Journal of Organizational Behavior, 35(3), 358–372.
Miller, E. K., & Cohen, J. D. (2001). An integrative theory of prefrontal cortex function. Annual Review of Neuroscience, 24, 167–202.
Morgan, G., Stephanou, A., & Simpson, B. (2000). Aptitude profile test series. Melbourne, Victoria: Australian Council for Educational Research.
Nohe, C., Michaelis, B., Menges, J. I., Zhang, Z., & Sonntag, K. (2013). Charisma and organizational change: A multilevel study of perceived charisma, commitment to
change, and team performance. The Leadership Quarterly, 24(2), 378–389.
Padilla, A., Hogan, R., & Kaiser, R. B. (2007). The toxic triangle: Destructive leaders, susceptible followers, and conducive environments. The Leadership Quarterly, 18, 176–194.
Parker, S. K., Bindl, U. K., & Wu, C. -H. (2013). Who is proactive and why? Unpacking individual differences in employee proactivity. Advances in positive organizational
Parker, S. K., Williams, H. M., & Turner, N. (2006). Modeling the antecedents of proactive behavior at work. Journal of Applied Psychology, 91(3), 636–652.
Perkins, A. M., & Corr, P. J. (2005). Canworriers bewinners? The association betweenworrying and job performance. Personality and Individual Differences, 38(1), 25–31.
Ployhart, R. E., Lim, B. C., & Chan, K. Y. (2001). Exploring relations between typical and maximum performance ratings and the five factor model of personality.
Personnel Psychology, 54(4), 809–843.
Podsakoff, P. M.,MacKenzie, S. B., Lee, J. Y., & Podsakoff, N. P. (2003). Common method biases in behavioral research: A critical review of the literature and recommended
remedies. Journal of Applied Psychology, 88(5), 879–903.
Podsakoff, P. M.,MacKenzie, S. B., Moorman, R. H., & Fetter, R. (1990). Transformational leader behaviors and their effects on followers’ trust in leader, satisfaction, and
organizational citizenship behaviors. Leadership Quarterly, 1(2), 107–142.
Posner, M. I., & Rothbart, M. K. (1998). Attention, self-regulation and consciousness. Philosophical Transactions of the Royal Society B Biological Sciences, 353(1377),
Posner, M. I., Rothbart, M. K., Sheese, B. E., & Tang, Y. (2007). The anterior cingulate gyrus and the mechanism of self-regulation. Cognitive, Affective, & Behavioral
Neuroscience, 7(4), 391–395.
Rafferty, A. E., & Griffin, M. A. (2004). Dimensions of transformational leadership: Conceptual and empirical extensions. The Leadership Quarterly, 15, 329–354.
Rothbart, M. K., Sheese, B. E., & Posner, M. I. (2007). Executive attention and effortful control: Linking temperament, brain networks, and genes. Child Development
Perspectives, 1(1), 2–7.
Schyns, B., & Schilling, J. (2013). How bad are the effects of bad leaders? A meta-analysis of destructive leadership and its outcomes. Leadership Quarterly, 24(1),
Shamosh, N. A., DeYoung, C. G., Green, A. E., Reis, D. L., Johnson,M. R., Conway, A. R. A., et al. (2008). Individual differences in delay discounting: Relation to intelligence,
working memory, and anterior prefrontal cortex. Psychological Science, 19(9), 904–911.
Shamosh, N. A., & Gray, J. R. (2007). The relation between fluid intelligence and self-regulatory depletion. Cognition and Emotion, 21(8), 1833–1843.
Sherman, W. S., Hitt, M. A., DeMarie, S.M., & Keats, B.W. (1999). Organizationalmorphing. The challenges of leading perpetually changing organizations in the twenty-first
century, Vol. 1.
Smillie, L. D., Yeo, G. B., Furnham, A. F., & Jackson, C. J. (2006). Benefits of all work and no play: The relationship between neuroticism and performance as a function of
resource allocation. Journal of Applied Psychology, 91(1), 139–155.
Spillane, R. (2012, February 21). Personality tests have too many deep-seated problems. The Australian Financial Review, 51.
Tepper, B. J. (2000). Consequences of abusive supervision. Academy of Management Journal, 43(2), 178–190.
Tepper, B. J. (2007). Abusive supervision in work organizations: Review, synthesis, and research agenda. Journal of Management, 33(3), 261–289.
Tett, R. P., & Guterman, H. A. (2000). Situation trait relevance, trait expression, and cross-situational consistency: Testing a principle of trait activation. Journal of
Research in Personality, 34(4), 397–423.
Tett, R. P., & Simonet, D. V. (2011). Faking in personality assessment: A “multisaturation” perspective on faking as performance. Human Performance, 24(4), 302–321.
Tsui, A. S., & Ashford, S. J. (1994). Adaptive self-regulation: A process view of managerial effectiveness. Journal of Management, 20(1), 93–121.
van Noordt, S. J., & Segalowitz, S. J. (2012). Performance monitoring and the medial prefrontal cortex: A review of individual differences and context effects as a window
on self-regulation. Frontiers in Human Neuroscience, 6, 197.
Vasilopoulos, N. L., Reilly, R. R., & Leaman, J. A. (2000). The influence of job familiarity and impression management on self-report measure scale scores and response
latencies. Journal of Applied Psychology, 85(1), 50–64.
Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of briefmeasures of positive and negative affect: The PANAS scales. Journal of Personality and
Social Psychology, 54(6), 1063–1070.
Wood, C. C., & Jennings, J. R. (1976). Speed–accuracy tradeoff functions in choice reaction time: Experimental designs and computational procedures. Perception &
Psychophysics, 19(1), 92–102.
Wu, C. H., & Wang, Y. (2011). Understanding proactive leadership. Advances in Global Leadership, 6(6), 299–314.
Yeow, J., & Martin, R. (2013). The role of self-regulation in developing leaders: A longitudinal field experiment. The Leadership Quarterly, 24(5), 625–637.
Zaccaro, S. J. (2007). Trait-based perspectives of leadership. American Psychologist, 62(1), 6–16.
Zaccaro, S. J. (2012). Individual differences and leadership: Contributions to a third tipping point. Leadership Quarterly, 23(4), 718–728.