Download as pdf or txt
Download as pdf or txt
You are on page 1of 42

r Academy of Management Annals

2023, Vol. 17, No. 1, 331–372.


https://doi.org/10.5465/annals.2020.0340

LEADERSHIP STYLES: A COMPREHENSIVE


ASSESSMENT AND WAY FORWARD
THOMAS FISCHER
University of Geneva

SIM B. SITKIN
Duke University

We systematically review eight positive (authentic, charismatic, consideration and initi-


ating structure, empowering, ethical, instrumental, servant, and transformational leader-
ship) and two negative leadership styles (abusive supervision and destructive leadership)
and identify valence-based conflation as a limitation common to all ten styles. This limi-
tation rests on specifying behaviors as inherently positive or negative and leads to mixing
the description of the content of leadership behaviors with the evaluation of their under-
lying intentions, quality of execution, or behavioral effects. We outline how this confla-
tion leads to amalgamation, construct redundancy, and most problematically, causal
indeterminacy, which calls into question the entire evidence base of leadership style
research. These weaknesses are not limited to the ten leadership styles but are inherent in
the valenced research logic that has been dominant for seventy years. Thus, the common
finding that positive leadership styles lead to positive outcomes and negative styles lead
to negative outcomes might be an artifact of conflation rather than a reflection of reality.
To address these concerns, we suggest distinguishing between intended and displayed
leadership styles, as well as their realized effects. We also call for utilizing a configura-
tional approach. These recommended actions would provide a strong foundation for
future research on leadership styles.

Leadership is an ever-growing field of research in Park, 2017). For instance, Lemoine et al. (2019: 177)
management and applied psychology. Leadership stated that their “comparative review of the three
scholars have noted that research has “made tremen- dominant moral approaches (i.e., authentic, ethical,
dous progress in uncovering some of the enduring and servant leadership) clearly indicates that moral
mysteries associated with leadership” (Avolio, leadership behaviors positively impact a host of
Walumbwa, & Weber, 2009: 442), including estab- desirable organizationally relevant outcomes.” Like-
lishing the widely shared consensus that “positive” wise, Tepper et al. (2017: 125) concluded that
styles of leadership cause desirable outcomes and “exposure to abusive supervision is associated with
“negative” styles of leadership cause undesirable a broad range of dysfunctional outcomes that are
outcomes. According to many leadership scholars, observed at both the individual and team levels.”
evidence has shown that people become more effec- The implications for practice would be significant if
tive when they engage in such positive styles as these conclusions rested on solid evidence—one
authentic, charismatic, empowering, ethical, instru- could just follow the “do-good logic” (which we
mental, servant, or transformational leadership, and define as the belief that engaging in “positive” styles is
less effective when they engage in such negative generally effective) and the “don’t-do-bad logic” (the
styles as abusive supervision or destructive leader- belief that engaging in “negative” styles is generally
ship (e.g., Cheong, Yammarino, Dionne, Spain, & ineffective) to achieve positive leadership outcomes.
Tsai, 2019; Krasikova, Green, & LeBreton, 2013; In contrast with this widespread and enthusiasti-
Lemoine, Hartnell, & Leroy, 2019; Tepper, Simon, & cally advocated conceptualization, an increasing
amount of scholarship has challenged popular leader-
We thank John Antonakis and J€ org Dietz for their valu- ship style constructs and the validity of their evidence
able feedback on early versions of the article. Correspon- base. For instance, Antonakis, Bendahan, Jacquart,
dence concerning this article should be directed to and Lalive (2010) identified large parts of empirical
Thomas Fischer. leadership research as being limited by statistical
331
Copyright of the Academy of Management, all rights reserved. Contents may not be copied, emailed, posted to a listserv, or otherwise transmitted without the copyright holder's express
written permission. Users may print, download, or email articles for individual use only.
332 Academy of Management Annals January

misspecifications, stating that such misspecifica- evaluation of the quality of execution or effects. Thus,
tions undermine the trustworthiness of research find- well-executed behaviors or behaviors with autonomy-
ings. On conceptual grounds, Van Knippenberg and enhancing and inspirational effects are more likely to
Sitkin (2013) delineated confusion in construct defini- be classified as instances of a positive leadership style
tion and measurement in research into charismatic than are behaviors that are poorly executed or have
and transactional-transformational leadership styles. negative effects.
More recently, similar critiques have been articulated We have found that all 10 reviewed leadership styles
for other leadership styles (e.g., Banks, Fischer, Gooty, study valenced and conflated bundles of behaviors
& Stock, 2021a; Fischer, Tian, Lee, & Hughes, 2021). that are supposed to constitute either good or bad lead-
Moreover, scathing critiques have called much leader- ership. This approach is problematic because it makes
ship research fundamentally tautological (Alvesson, the research prone to conflating leadership behaviors
2020; Alvesson & Einola, 2019) and its evidence “BS” with underlying intentions, quality of execution, and
(i.e., bullshit) because it fails to accurately represent realized effects. To overcome these concerns and to
organizational reality (Pfeffer, 2015). advance leadership style research, we define and call
In our systematic review, we map the research for disentangling intended and displayed leadership
landscape and systematically assess the conceptual styles as well as their realized effects.
and empirical validity of 10 of the most influential We use the term valenced leadership styles in a
leadership style research programs. We assess the broad sense to refer to patterns of leadership behaviors
leadership styles using a framework that distinguishes that are specified as good or bad. The basis for judging
four behavioral aspects: behavioral content (“what a a behavior’s goodness or badness can be based on
leader does”), underlying intentions (“why the leader whether presumed underlying intentions, quality of
does it”), quality of execution (“how well they do it”), execution, or behavioral effects are good or bad. This
and effects (“how effective” in terms of contributing to goodness or badness can manifest, for instance, in
organizational performance, normative achievements effectiveness (e.g., contribution to organizational effec-
or compliance or other outcomes). The first behavioral tiveness) or normativity (e.g., normative compliance or
aspect is descriptive (i.e., behavioral content), while normatively laudable achievements). Valence-based
the other three aspects are evaluative. We find that conflation refers to situations in which otherwise quite
the definitions and measures of leadership styles fre- distinct behaviors are clustered and labeled as being
quently conflate two or more of the four aspects, nota- part of a style simply because of shared valence. This
bly behavioral content with at least one evaluative issue is illustrated in the case of the ethical leadership
aspect. More generally, leadership styles are com- style, as defined by Brown, Trevi~ no, and Harrison
monly studied as patterns of informal and interactive (2005), which includes normatively laudable inten-
behaviors that are believed to foster certain desir- tions, effective execution of leadership behaviors,
able or undesirable follower-, team-, or organization- and the realization of normatively desirable effects.
related objectives or consequences. However, The main commonality among the elements of this
asserting that a pattern of behaviors (i.e., a leadership leadership style is that all of the elements are deemed
style) exists simply because they include all-positive positive.
or all-negative attributes is problematic. This problem Building upon our systematic review, we can
is exacerbated when the asserted pattern of behaviors explain how valence-based conflation increases
includes not only behaviors but also intentions, out- the risk of amalgamation and construct redun-
comes, and assessments of the quality (e.g., execution dancy in the conceptualization of leadership styles
excellence or effectiveness) of the behaviors. Items and consistently leads to causal indeterminacy.
from academically popular questionnaires of leader- First, valence-based conflation makes leadership
ship styles serve as an illustration. For instance, style constructs prone to amalgamation in the sense
whether a leader acts with the primary intention of that they are bundling behaviors that are conceptu-
promoting ethical conduct (i.e., ethical leadership) ally unrelated except for shared valences. Constructs
cannot be judged solely based on behavioral descrip- involving valenced leadership styles tend to be amal-
tions but requires an evaluation of positive intentions gamated because they lack a clear conceptual ratio-
too. Similarly, whether a leader provides autonomy nale for bundling behaviors, and bundle behaviors
from bureaucratic constraints (i.e., empowering lead- merely based on whether intent, quality of execu-
ership) or is intellectually stimulating (i.e., transfor- tion, or effects are positive or negative.
mational leadership) cannot be judged solely based Second, valence-based conflation makes leader-
on behavioral descriptions either but also requires an ship styles prone to construct redundancy in the
2023 Fischer and Sitkin 333

sense that these styles overlap not only in terms of circumstances—their hard-hitting behaviors are
describing leadership behaviors (see Banks, Gooty, likely evaluated as having positive transformative
Ross, Williams, & Harrington, 2018) but also in posi- effects. If they fail, though—due to, for instance,
tive or negative evaluations of intent, execution, resisting employees—their hard-hitting behaviors
and effects. For instance, both empowering and ser- are likely to be evaluated as having abusive effects.
vant leadership overlap in positive evaluations, In the first case, positive transformative effects coin-
such that both leadership styles produce the effect cide with success, and in the second case, negative
of increased follower engagement and participation abusive effects coincide with failure; transforma-
at work. tional leadership seems to cause success and abusive
Third, and most problematically, valence-based supervision seems to cause failure. The first case
conflation necessarily leads to causal indeterminacy. affirms the do-good logic of leadership, and the sec-
We define causal indeterminacy as the impossibility ond case affirms the don’t-do-bad logic of leadership.
of determining whether a leadership style–outcome However, both cases rest on identical content of
link is due to leadership behaviors or other factors. leadership behaviors. There is only variation in the
Valenced leadership styles are causally indeterminate
evaluative aspects of leadership that have also non-
because they conflate behavioral content with under-
behavioral antecedents, and these evaluative aspects
lying intentions, quality of execution, or behavioral
account for a causally indeterminate link between
effects. Consequently, it is unclear whether empirical
findings for positive and negative leadership styles leadership behaviors and effects.
rest on a behavioral grounding, or if there are alterna- Our systematic review and the assessment show
tive explanations for style–outcome relationships. The how valence-based conflation leads to amalgam-
entire existing evidence base claiming that positive ation, construct redundancy, and causal indetermi-
leadership styles cause positive outcomes and negative nacy. After the systematic review, we broaden the
leadership styles cause negative outcomes might perspective beyond the 10 leadership styles and
be erroneous, and empirical findings supporting assess the field overall by tracing the evolution and
the do-good logic and the don’t-do-bad logic of basic underlying logic of leadership style research
leadership styles might be artifacts of causal inde- that has led to the problems found in each of the
terminacy. For instance, a positive link between leadership styles we examine. We find that over
transformational leadership and positive outcomes seven decades valence-based conflation has mani-
(e.g., follower performance, satisfaction) and a nega- fested in four different sublogics, but that the over-
tive link between abusive supervision and negative arching logic of valenced specifications of leadership
focal outcomes (e.g., stress, turnover) is causally inde- styles are a lasting characteristic of the whole field
terminate because transformational leadership and of research. In addition, causal indeterminacy is also
abusive supervision conflate leadership behaviors a constant and inherent part of the whole field,
with their effects (see also Fischer et al., 2021; Van given that valence-based conflation leads to causal
Knippenberg & Sitkin, 2013). indeterminacy.
This causal indeterminacy can happen in two Based on this insight, we call for rethinking—
ways: (a) if the focal outcome is defined by transfor- but not abandoning—leadership style research.
mational or abusive effects, there is a tautology; (b)
For rethinking leadership style research, we outline
even when the outcome is not defined in transforma-
in the final section of our article how distinguishing
tional or abusive terms, a style–outcome link could
intended and displayed leadership styles, as well as
be mistaken for a causal effect when the focal out-
their realized effects, could help avoid valence-based
come (e.g., follower performance) is caused by or
is cooccurring with another outcome (e.g., posited conflation and causal indeterminacy. In addition, we
transformational or abusive effects, which are con- argue why such deconflation reduces the risk of amal-
flated aspects of leadership style constructs) rather gamation and construct redundancy. Moreover, we
than being caused by leader behaviors (i.e., behav- sketch how studying leadership styles configuration-
ioral content). Take the example of a manager with ally further contributes to raising conceptual clarity
a hard-hitting behavior targeted at employees to in leadership style research and, in this way, contrib-
improve customer service processes (e.g., by pushing utes to reducing amalgamation and construct redun-
for process improvements with an aggressive voice). dancy. Put differently, there is a need to go beyond
If the manager is successful in implementing leadership styles as valenced catch-all concepts in
their change initiative—also, for instance, due to order to avoid conflation while still ensuring concep-
receptive employees and favorable organizational tual integration.
334 Academy of Management Annals January

PLAN FOR THE SYSTEMATIC REVIEW: review shows, also conflate the content of behaviors
RATIONALE, SCOPE, AND APPROACH with good quality of execution, and, to some degree,
even good intentions. Constructs and labels change,
Rationale of the Review
but the research logic and its shortcomings remain
In our broad, meta-level review of 10 leadership the same. Our meta-level review of multiple leader-
styles that academics often study, we pay particular ship styles and the assessment of their common
attention to the risk of valence-based conflation and research logic and common limitations avoids per-
the resulting risks of amalgamation, construct redun- petuating the same limitations in another disguise
dancy, and causal indeterminacy. We believe that and allows us to outline a way forward. Such a meta-
such a meta-level review is more appropriate as a level review builds on previous reviews but goes
next step for addressing the limitations of conflation, beyond and extends them.
amalgamation, construct redundancy, and causal First, our work uncovers valence-based conflation
indeterminacy when compared with yet another the- as a common weakness across research on leadership
ory article, commentary, or review of a single leader- styles. This finding synthesizes previous review-based
ship style. Our call for a fundamental reorientation critiques and expands their relevance for all leadership
of research practice is admittedly a very strong claim, styles. For instance, Van Knippenberg and Sitkin
which is why we have documented the prevalence (2013) lamented that charismatic and transactional-
of the limitations for all 10 leadership styles. Such a transformational leadership confound behaviors with
wide claim and review is barely feasible within the effects. Alvesson and Einola (2019) noted that there is
scope of a traditional theory article or commentary. “excessive positivity” in the definition and measure-
However, because we believe such a strong reorien- ment of authentic leadership, and Banks et al. (2021a)
tation is necessary, our meta-level review is designed and Fischer et al. (2021) found that there is conflation
to carefully document these limitations for the whole of leader behaviors with positive or negative follower
field of leadership style research. evaluations for ethical leadership and abusive supervi-
In addition, reviews of a single leadership style sion, respectively. All these previous reviews have
have already been conducted and have documented pointed to some form of valence-based conflation.
conflation, amalgamation, construct redundancy, and Extending such work, we distinguish three types of
causal indeterminacy (e.g., Banks et al., 2018; Fischer valence-based conflation—namely behavior–intent,
et al., 2021; Hoch, Bommer, Dulebohn, & Wu, 2018; behavior–execution, and behavior–effect conflation—
Van Knippenberg & Sitkin, 2013), but have missed and we show that all reviewed leadership styles are
the overall pattern. In other words, the focus on single limited by at least one such type of conflation. We con-
style assessments and the absence of a broad, meta- clude that valence-based conflation is a defining fea-
level review has had the consequence of problem per- ture of leadership style research to date.
petuation because the limitations and shortcomings Second, we show that valence-based conflation
identified for one or a few leadership styles have been makes leadership style constructs prone to amalgam-
unknowingly replicated in research that has simply ation. Yukl (1999) lamented that there is ambiguity
moved on to another leadership style with the very in whether a behavior is charismatic or transforma-
same issues. For instance, Yukl (1999) criticized ambi- tional, and Van Knippenberg and Sitkin (2013)
guity in whether a behavior is transformational, and highlighted the absence of causal models specify-
Van Knippenberg and Sitkin (2013) criticized transfor- ing how different dimensions of transformational
mational leadership for confounds between behaviors
leadership relate to each other to represent a single,
and their hypothesized effects. Following these cri-
coherent construct. While these works identified
tiques, scholars moved to studying alternative leader-
amalgamation for charismatic and transactional-
ship styles such as authentic or ethical leadership
(Lemoine et al., 2019). At first glance, the shift in transformational leadership, we extend their in-
scholarly activity could be interpreted as indicating sights and sketch how valence makes all reviewed
that researchers have learnt from the two critiques and leadership styles—actually all possible valenced
started using better concepts. With a closer look, conceptualizations of leadership styles—prone to
however, we have to acknowledge that authentic amalgamation. Thus, we concur with and go beyond
and ethical leadership in fact replicate the previ- Carton’s (2022) review, which indicated that lead-
ously identified issues as they continue to confound ership constructs lump together multiple distinct
behaviors with their effects (see Alvesson & Einola, behaviors. Our review shows that valenced specifi-
2019; Banks et al., 2021a) and, as our systematic cations of leadership are the source of grouping
2023 Fischer and Sitkin 335

conceptually unrelated behaviors into amalgams. Scope of the Review


Future work needs to avoid valenced specifications
For each of the 10 leadership styles we review—
of behaviors.
eight positive and two negative ones—we searched in
Third, narrative reviews (e.g., Anderson & Sun,
Web of Science for articles published between 2000
2017; Banks et al., 2018), as well as meta-analyses
and 2020 in widely cited journals publishing such
(e.g., Banks, McCauley, Gardner, & Guler, 2016; research.1 We chose the year 2000 because after Yukl’s
Hoch et al., 2018), have identified construct redun- (1999) critique of charismatic and transactional-
dancy in leadership styles. Our review uncovers transformational leadership, scholars suggested a
valence-based conflation as a major source of this number of alternative styles, including some of
previously identified construct redundancy. those covered in this review.
Fourth, we note that conceptual conflation leads For each leadership style, we review 20 articles
to causal indeterminacy in the relationship between that reflect the state of the science.2 Inclusion of
leadership styles and focal outcomes. Given the articles was based on the following criteria in
identified conflation problems, we show that it is descending order of priority: (a) review articles, (b)
impossible to infer the effects of leadership behav- articles marked as highly cited in Web of Science
iors on outcomes (e.g., follower performance or satis- (i.e., top 1% per year), and (c) articles with the highest
faction). In the case of behavior–effect conflation, absolute number of citations (for more details, see
causal indeterminacy is obvious because one out- Appendix A). Including review articles allows us
come (e.g., follower performance) is explained by to integrate insights from previous scholarly work,
another outcome (e.g., empowering or transforma- while including highly cited articles allows us to
tive effects), instead of explaining an outcome by review the most influential work that has shaped
its causes (e.g., nonconflated content of behaviors). research on leadership styles while still keeping
Causal indeterminacy is also apparent in the case of the sample size manageable for a narrative review.
behavior–execution conflation, which characterizes Together, this approach yielded a set of 17 concep-
all 10 reviewed styles. Evaluations of the quality of tual, 87 empirical, and 31 review articles. In addition,
executing leadership behaviors are sometimes also we included 22 articles that met either of the two cri-
called leadership performance (cf. Campbell, 2012; teria: (a) review articles, meta-analyses, or critical
Campbell & Wiernik, 2015). That is, positive leader- assessments published outside of our review period;
ship styles are specified by good performance (i.e., or (b) articles suggesting a definition or measure of the
high quality execution) and negative leadership styles leadership style that other articles in our review have
by poor performance (i.e., poor quality execution). It built on. Examples are the conceptualizations of char-
is causally indeterminate to say that good or bad per- ismatic leadership by Antonakis, Bastardoz, Jacquart,
formance is related with good or bad outcomes. In and Shamir (2016) and empowering leadership by
addition, even for behavior–intent conflation, causal Kirkman and Rosen (1999).
indeterminacy is present. While intentions are an
antecedent of behaviors, intentions are unobservable. 1
Only a leader’s claims about leadership intentions Widely cited journals in management and applied
psychology that were included in the review are: Academy
and (other) people’s expressions of attributions
of Management Annals, Academy of Management Journal,
concerning such intentions are observable. Making Academy of Management Review, Administrative Science
one’s intentions explicit is a leadership behavior, Quarterly, Annual Review of Organizational Psychology
whereas attributions concerning leadership inten- and Organizational Behavior, Annual Review of Psychol-
tions are partially leadership outcomes (Martinko ogy, Journal of Applied Psychology, Journal of Manage-
et al., 2018; Meindl, Ehrlich, & Dukerich, 1985). That ment, Leadership Quarterly, Organization Science, and
is, insofar as leadership intentions are captured by Strategic Management Journal.
2
attributions, behavior–intent conflation specifies Exceptions are consideration-initiating structure lead-
leadership by its outcomes too. Because all valenced ership, destructive leadership, and instrumental leader-
ship, because there are fewer than 20 articles fitting the
leadership styles are specified by outcomes of lead-
search criteria. In case of consideration and initiating
ership, it is unsurprising to note that good leadership
structure leadership, the likely reason for the lower num-
styles seemingly relate to good outcomes, and bad ber of articles is the relatively old age of these concepts
leadership styles relate to bad outcomes. Yet such a because much research was conducted before the 2000s. It
do-good and don’t-do-bad logic of leadership rests is the opposite for instrumental leadership as a relatively
on causally indeterminate grounds. new concept.
336 Academy of Management Annals January

Approach of the Review and the Four Behavioral that the evaluative component is also called (leader-
Aspects Framework ship) performance [see Campbell, 2012; Campbell
& Wiernik, 2015]). Fourth, scholars have already
To assess valence-based conflation in leadership
warned about conflating leadership behaviors with
style research, we utilized a self-developed frame-
their effects (e.g., Banks et al., 2021b; Van Knippen-
work that we call the four aspects of leadership
behaviors. As we noted in the introduction to this berg & Sitkin, 2013). The four aspects framework is
article, the four aspects framework distinguishes one also similar to the typology of prosociality at work by
descriptive and three evaluative aspects of leader- Bolino and Grant (2016) that distinguishes prosocial
ship behaviors: behavioral content (“what a leader motives, prosocial behaviors, and prosocial effects.
does”), underlying intentions (“why the leader does Our framework refers to intentions instead of motives,
it”), quality of executing behaviors (“how well they and, more substantively and in line with industrial-
do it”), and behavioral effects (“how effective” in organizational psychology, we further split behavioral
terms of contributing to organizational performance content and quality of behavioral execution (Borman &
or normative achievements or compliance). Brush, 1993; Campbell & Wiernik, 2015). As such,
This framework is grounded in existing research. each of the components of the four aspects framework
First, studying behavioral content has commonly have been already discussed extensively in leader-
been regarded as the heart of leadership research ship or organizational behavior research, and we have
(Banks, Woznyj, & Mansfield, 2021b; Yukl, 2012). drawn them together in one framework to assess lead-
Second, leader intentions and attributions of such ership style research more systematically. We sum-
intentions are a central part of the longstanding attri- marize the four aspects framework in Table 1.
butional school of leadership (e.g., Martinko, Harvey, The four aspects framework allows us to assess
& Douglas, 2007; Meindl et al., 1985). Third, the qual- valence-based conflation because three of the four
ity of executing behaviors refers to “how well” a per- aspects—intentions of behaviors, the quality of
son performs a behavior. The distinction between executing behaviors, and effects of behaviors (i.e.,
content and quality of executing behaviors resembles except behavioral content)—are evaluative and point
the distinction that industrial-organizational psychol- to goodness or badness. See Table 2 for more details
ogists make between the descriptive (“what”) and about the evaluative aspects of the framework and
evaluative (“how well”) component of behavior (note how they lead to conflation.

TABLE 1
The Four Aspects of Leadership Behaviors
Descriptive Aspect Evaluative Aspects
Intent Quality Outcome
Content “Why” a leader does what “How well” a leader “How effective” the
“What” a leader does they do executes the behavior behavior is

Construct space Content of behaviors Intentions underlying Quality of execution Effectiveness with
behaviors respect to relevant
and normatively
acceptable goals
Representation of the Observation of Unobservable (often assessed Assessment of behaviors Indicators of
construct behaviors with a proxy such as with respect to quality outcomes or
attribution of intentions, standards achievements
which is a judgment)
Potentially underlying Mix of person and Leader character Leader competency Leader success
root cause situation
Example: Giving Content (e.g., use (Self-)Attribution of Quality of execution (e.g., Effects (e.g., feedback
feedback or number of intention (e.g., feedback clarity and specificity that is helpful or
laudatory words in to help or harm the of feedback) harmful for follower
the feedback talk) follower’s development) development)

Notes: Behavioral content and quality of execution are the descriptive and evaluative aspects of behaviors, respectively. Underlying
intentions and behavioral effects are further aspects related to behaviors, but obviously not part of behaviors as such.
2023 Fischer and Sitkin 337

TABLE 2
An Overview of the Three Subtypes of Valence-Based Conflation
Illustration for
Definition of Conflation Implications for Interpreting Empowering Leadership
Subtype Causal Links (EmpL)

Behavior–intent Classifying behaviors by Indeterminate whether If EmpL presupposes the


conflation positive or negative behaviors or (attributions intention to foster
underlying intentions of) underlying intentions autonomy
(“why”) account for the causal link
Behavior–execution Classifying behaviors by the Indeterminate whether EmpL presupposes good
conflation degree to which their content of the behavior or execution of behaviors
execution is of high quality its competent execution for fostering autonomy
(“how well”) accounts for the causal link
Behavior–effect Classifying behaviors by Tautological because the EmpL presupposes
conflation positive or negative effects behavior (as the cause) is increased levels of
(“how effective or moral”) defined by its effect follower autonomy

Notes: We identify valence-based conflation of leadership styles with the four-aspects framework of leadership behaviors such that
mixing the descriptive aspect (i.e., content) with at least one evaluative aspect (i.e., intention, quality of execution, effects; see also Figure 1)
constitutes conflation.

MAPPING THE TERRAIN AND ASSESSING Consideration and Initiating Structure Leadership
CONFLATION: A SYSTEMATIC REVIEW OF 10
Conceptual evolution. The root concepts of mod-
LEADERSHIP STYLES
ern leadership style research are consideration and
In this section, we map, review, and assess the initiating structure leadership, which basically
10 leadership styles across the four behavioral refers to relationship-oriented (“consideration”)
aspects of leadership framework (see also Table 1). and task-oriented (“initiating structure”) leadership
For each leadership style, we start by assessing its behaviors. Conceptualization and measurement of
conceptual evolution. Then, we revisit the domi- these styles date back to the Ohio State Leadership
nant conceptualization of the leadership style. Next, Studies that aspired to describe leader behaviors rel-
we map the dominant conceptualization according evant for problem solving in organizations (Fleish-
to the four aspects framework to assess whether there man, 1953; Fleishman, Mumford, Zaccaro, Levin,
is valence-based conflation. Subsequently, if appli- Korotkin, & Hein, 1991). This research provided an
cable, we assess alternative views or review previous academic foundation for leadership frameworks that
critiques of the conceptualization of that style. are popular in leadership training, such as Blake and
Finally, we summarize key insights derived from Mouton’s (1964) managerial grid, which suggests a
reviewing the leadership style, which highlights a combination of high relations- and task-orientation,
nascent deconflated view of leadership. We summa- and Hersey and Blanchard’s situational leadership
rize key findings in Table 3. model, which suggests that leaders should vary their
We start our systematic review with consideration display of relations- and task-oriented behaviors,
and initiating structure leadership and transforma- contingent on the follower’s maturity level (Blan-
tional leadership because the conceptualization and chard, Zigarmi, & Nelson, 1993).
measurement of these two approaches served as Notwithstanding the impact of research into consid-
starting points for the definition of other well- eration and initiating structure leadership, in recent
studied leadership styles that are part of our system- decades research attention shifted toward other con-
atic review. We continue with authentic, empower-
structs. Accordingly, Judge et al. (2004: 36) referred to
ing, ethical, and servant leadership (in alphabetical
consideration and initiating structure as the “forgotten
order) as typical positive leadership styles. Then, we
ones,” even though scholars reported meta-analytic
review charismatic and instrumental leadership,
support for their effectiveness (e.g., DeRue, Nahrgang,
which share key features of positive leadership
Wellman, & Humphrey, 2011a; Judge et al., 2004).
styles yet are characterized by more idiosyncrasies
in conceptualization. Finally, we review abusive Reviewing the dominant conceptualization. Con-
supervision and destructive leadership as negative sideration is, according to Judge et al. (2004: 36) “the
leadership styles. degree to which a leader shows concern and respect
338 Academy of Management Annals January

TABLE 3
Mapping the Terrain of Contemporary Leadership Style Research
Behavioral Aspects of Leadership Behaviorsa
Leadership Dominant Definition or Quality of
style Dimensions Content Intention execution Effect Classical Reference(s)

Consideration-initiating (1) Consideration   () Stogdill (1962), Judge,


structure leadership (2) Initiating structure   () Piccolo, and Ilies
(2004)
Transformational (1) Idealized influence    Bass (1985), Bass and
leadership (attributes and behaviors) Avolio (1995)
(2) Inspirational motivation   
(3) Intellectual stimulation   
(4) Individualized consideration   ()
Authentic leadership (1) Self-awareness () () Walumbwa, Avolio,
(2) Relational transparency   Gardner, Wernsing,
(3) Internalized moral () () and Peterson (2008)
perspective
(4) Balanced processing  
Empowering leadership (1) Enhancing the   Ahearne et al. (2005)
meaningfulness of work
(2) Fostering participation in   
decision making
(3) Expressing confidence in  
high performance
(4) Providing autonomy from   
bureaucratic constraints
Ethical leadership Demonstrating and promoting     Brown et al. (2005)
normatively appropriate
conduct
Servant leadership (1) Conceptual skills  Liden, Wayne, Zhao, and
(2) Empowerment    Henderson (2008)
(3) Helping subordinates grow  ()  
and succeed
(4) Putting subordinates first  
(5) Behaving ethically    
(6) Emotional healing  ()  
(7) Creating value for the ()  
community
Charismatic leadership Charisma is a subconstruct of    Bass (1985), Bass and
transformational leadership Avolio (1995)
Instrumental leadership (1) Environmental monitoring   Antonakis and House
(2) Strategy formulation and   () (2014)
implementation
(3) Path–goal facilitation   
(4) Outcome monitoring   
Abusive supervision Sustained display of hostile,  ()   Tepper (2000)
nonphysical behavior
Destructive leadership Systematic and repeated    Einarsen, Aasland, and
behavior that violates the Skogstad (2007)
legitimate interest of the
organization by undermining
or sabotaging subordinate- and
organization-level factors and
outcomes

Notes: The table shows that all leadership styles and their single dimensions are valenced because they cover intentions or quality of
execution or effects, and thus do not objectively describe the behavioral content of leadership behaviors.
a
A tick in parentheses indicates that the aspect of leadership is only present in the measurement tool but not the conceptualization.
2023 Fischer and Sitkin 339

for followers, looks out for their welfare, and expresses positive judgments about a leader, and that are con-
appreciation and support.” In contrast, initiating ceptually closer to effects of leadership behaviors
structure is “the degree to which a leader defines and than to being leadership behaviors themselves, are
organizes his role and the roles of followers, is ori- thus mistakenly used to measure leadership behav-
ented toward goal attainment, and establishes well- iors; they do not measure what they are supposed to
defined patterns and channels of communication” measure. In fact, some items cannot be characterized
(Judge et al., 2004: 36). The most popular operation- accurately as behavior descriptions because they
alization of consideration and initiating structure refer to unobservable characteristics or intentions
leadership is the Leadership Behavior Description (“speaks from a strong inner conviction”) or attitudes
Questionnaire (LBDQ) in its version LBDQ-XII (Judge (“remains calm when uncertain about coming even-
et al., 2004; Stogdill, 1962). Judge et al. (2004) and ts”). In addition, several items confound behaviors
DeRue et al. (2011a) synthesized decades of research with their effects (e.g., “arguments are convincing,”
on consideration and initiating structure leadership “settles conflicts in the group when they occur”). For
and offered meta-analytic support for its effective- example, LBDQ-XII fails to fully describe the use of
ness. More recently, Keller (2006), Lambert, Tepper, leader behaviors in part because it mixes behavioral
Carr, Holt, and Barelka (2012), and Neubert, Kacmar, description with evaluation of the quality of the execu-
Carlson, Chonko, and Roberts (2008) found positive tion of those behaviors. While the definitions of consid-
effects as well. eration and initiating structure are only limited by
Mapping the dominant conceptualization accord- behavior– execution conflation, measurement items
ing to the four aspects framework to assess valence- are also limited by behavior–effect and behavior–
based conflation. The definition of initiating struc- intent conflation.
ture behaviors explicitly spans two behavioral aspects
Reviewing previous critiques. Generally, the valid-
of leadership: behavioral content (“defines and organ-
ity of the meta-analytic findings concerning this lead-
izes his role and the role of followers” [Judge et al.,
ership style is conditional on the conceptual and
2004: 36]) and quality of execution (“establishes well-
empirical rigor of the LBDQ. This rigor has been ques-
defined patterns and channels of communication”
tioned on psychometric grounds (Schriesheim &
[Judge et al., 2004: 36]). Thus, there is behavior–
Kerr, 1974). Perhaps even more importantly, Lord,
execution conflation. The definition of consideration
Binning, Rush, and Thomas (1978) found that
behaviors is a less clear-cut case of valence-based con-
responses to the LBDQ depend on performance infor-
flation. Yet if behaviors are more likely to be classified
mation such that high-performing leaders are reported
as showing concern or respect for followers when they
to score higher on consideration and initiating struc-
are executed well or produce positive effects (e.g., fol-
ture, irrespective of actual behaviors. Because per-
lower’s feeling of respect), then such classifications
formance information can influence evaluations of
rest on behavior–execution and behavior–effect con-
execution quality and effectiveness and do not reflect
flation too.
the impact of behavioral content, this finding provides
Indeed, behavior–execution and behavior–effect
additional support for our argument that consider-
conflation characterize LBDQ-XII (Judge et al., 2004;
ation and initiating structure are limited by valence-
Stogdill, 1962). For instance, “makes accurate deci-
based conflation.
sions” in LBDQ-XII rests as much on an evaluation of
accuracy as it rests on the description of decision- Summary. The formulation of leadership in terms
making. The qualifier “accurate” leads to conflating of consideration and initiating structure is a modern
behavioral content with the quality of executing classic in leadership style research; task-oriented
these behaviors—thus, scholars who develop mea- (“initiating structure”) and relation-oriented
surement tools should abstain from using such quali- (“consideration”) behaviors are key parts of leader-
fiers. Likewise, “persuades others that his ideas are ship research. Unfortunately, both the definition and
to their advantage” refers to persuasion behaviors the measurement of the two dimensions are inher-
but also evaluations of their effectiveness. Further- ently valenced. Because behaviors are specified by
more, some items are only evaluative, even though their contribution to follower and work group func-
they are part of a measure that is supposed to tioning—that is, behavioral content is only classified
describe leader behaviors. Illustrative items are “can as consideration and initiating structure behaviors if
reduce a madhouse to system and order” and “is they foster respective goals—behavior–execution and
friendly and approachable.” Items that capture behavior–effect conflation is the consequence.
340 Academy of Management Annals January

Transformational Leadership and is thus limited by both behavior–execution and


behavior–effect conflation. The original measure and
Conceptual evolution. The concept of transforma-
its refinements (e.g., Avolio, Bass, & Jung, 1999; Bass &
tional leadership dates back to the work of the politi-
Avolio, 1995) have generally found meta-analytic sup-
cal scientist Burns (1978). In management research,
port for leading to a variety of positive outcomes (Judge
Bass (1985) suggested conceptualizations and mea-
& Piccolo, 2004), but have not taken valence-based
surement of transformational leadership that have
conflation into account.
become classics in the field. The basic idea of trans-
Some scholarly work has gone beyond positivity
formational leadership is that leaders should not
to study the downsides of transformational leader-
limit themselves to fair relations-oriented and useful
ship (Kark, Shamir, & Chen, 2003; Lin, Scott, &
task-oriented behaviors with followers, which would Matta, 2019). However, these approaches have still
represent transactional leadership (and also resemble defined transformational leaders in terms of their
consideration and initiating structure behaviors). effects (i.e., based on the degree to which they are
Instead, Bass (1985) proposed that transformational influential on followers). As such, research on the
leaders appeal to values to induce performance that downside of transformational leadership has consid-
goes beyond expectations. Thus, transformational ered negative effects of highly influential leaders (cf.
leadership research has put people’s values center- the distinction between authentic and pseudo trans-
stage in leadership research, and is probably the most formational leadership [Bass & Steidlmeier, 1999])
highly cited stream of research in the field. More but has remained positively valenced by conflating
recently, though, research on transformational the content of behaviors with high-quality execution
leadership has declined, following Van Knippenberg of these behaviors and social influence effects on
and Sitkin’s (2013) critique of its conceptual and followers.
measurement-based underpinnings. Reviewing and assessing alternative conceptual-
Reviewing the dominant conceptualization. In izations according to the four aspects framework.
its currently most popular version, transformational Podsakoff, MacKenzie, Moorman, and Fetter (1990)
leadership rests on five dimensions: idealized introduced an alternative conceptualization of trans-
influence attributes, idealized influence behav- formational leadership that rests on the dimensions
iors, inspirational motivation, intellectual stimu- of vision articulation, providing an appropriate
lation, and individualized consideration (Bass & model, fostering acceptance of goals, setting high
Avolio, 1995). Typically, transformational leadership performance expectations, providing individualized
has been studied in conjunction with and distin- support, and offering intellectual stimulation. This
guished from the three-dimensional transactional lead- version of transformational leadership shares two
ership construct plus the one-dimensional laissez-faire dimensions with the Bass (1985) model: individual-
leadership construct, which together constitute the ized support and intellectual stimulation. In this
so-called “full range of leadership” (Antonakis, Avo- way, the models by Bass (1985) and Podsakoff et al.
lio, & Sivasubramaniam, 2003). (1990) share valence-based conflation inherent in
Mapping the dominant conceptualization accord- these dimensions. In addition, the dimension of fos-
ing to the four aspects framework to assess valence- tering acceptance of goals conflates behavioral con-
based conflation. When assessing transformational tent with high-quality execution and positive effects
leadership using the four aspects framework, we in fostering goals. The questionnaire items of Pod-
find that the construct not only includes behavioral sakoff et al. (1990) also reflect these conflated con-
content but also includes effects and points to high ceptual dimensions.
quality of execution. For instance, intellectual stimu- Reviewing previous critiques. The conceptualiza-
lation specifies a positive effect on followers, and the tion of transactional-transformational leadership as
successful realization of such a positive effect makes the “full-range model” sounds comprehensive, but
it more likely that the behavior was executed well. our review is consistent with Van Knippenberg and
The most popular measure of transformational Sitkin’s (2013) critique that the dimensions of trans-
leadership is the Multifactor Leadership Question- formational leadership confound behaviors with
naire (MLQ), which has been broadly applied in the their effects. Consequently, there is inherent confla-
studies reviewed here. There are different versions of tion, which undermines confidence in the empirical
the MLQ, but MLQ Form X is the most widely used evidence of primary studies (e.g., Gong, Huang, &
one (see Antonakis et al., 2003). The instrument Farh, 2009; Wang, Law, Hackett, Wang, & Chen,
reflects the positive valence of its underlying construct 2005) and meta-analyses of transformational
2023 Fischer and Sitkin 341

leadership (Judge & Piccolo, 2004). A recent review upon and promotes both positive psychological
by Siangchokyoo, Klinger, and Campion (2020) capacities and a positive ethical climate, to foster
raised some of these concerns too, suggesting that greater self-awareness, an internalized moral per-
transformational leadership research rests on prob- spective, balanced processing of information, and
lematic conceptual underpinnings. relational transparency on the part of leaders working
Summary. Transformational leadership is argu- with followers, fostering positive self-development.”
ably the most widely studied leadership construct. This definition is four-dimensional, including leader
The values-based notion that transformational lead- self-awareness, relational transparency, internalized
ers go beyond relations and tasks is appealing. How- moral perspective, and balanced processing. The
ever, both the review by Van Knippenberg and dimensionality is also reflected in the measurement
Sitkin (2013) and the more recent review by Siang- items of the Authentic Leadership Questionnaire
chokyoo et al. (2020) identified important flaws in (ALQ), and a recent review and a meta-analysis
the conceptualization and operationalization of the (Hoch et al., 2018; Lemoine et al., 2019) have argued
transformational leadership construct. Our review that authentic leadership leads to a host of posi-
points to valence-based conflation as accounting for tive outcomes.
these previously noted flaws. A positive trend, how- Mapping the dominant conceptualization ac-
ever, is that research on the flawed construct seems cording to the four aspects framework to assess
to be waning. Only one empirical paper published valence-based conflation. Two of the four authentic
after Van Knippenberg and Sitkin’s (2013) critique of leadership dimensions—leader self-awareness and
transformational leadership—that by Lin et al. internalized moral perspective—are cognitive and
(2019)—has been influential enough to qualify for hence do not clearly map onto any of the four behav-
inclusion in our sample for the review. ioral aspects of leadership. In contrast, the other two
dimensions of authentic leadership—balanced proc-
Authentic Leadership essing and relational transparency—describe behav-
ioral content (i.e., information processing, relational
Evolution of the concept. Authentic leadership activity) and also evaluate how well the behavior is
has been put forward to distinguish sincere and executed (i.e., balanced, transparency). Therefore, two
pseudo forms of transformational leadership (Bass dimensions of authentic leadership are nonbehavioral
& Steidlmeier, 1999). The notion of authenticity and two dimensions map onto multiple aspects of
quickly gained popularity among researchers and leadership and exhibit behavior–execution conflation.
helped put the importance of ethics center-stage Conflation also characterizes the ALQ-based mea-
in leadership research (Gardner, Cogliser, Davis, & surement of all four dimensions of authentic leader-
Dickens, 2011). The conceptual evolution of authen- ship, including for those dimensions that were defined
tic leadership research can be broken down in three nonbehaviorally. For instance, there is behavior–
phases. First, there was a phase of conceptual hetero- intent conflation in measuring the dimension of self-
geneity, in which authentic leadership was seen as a awareness (e.g., “says exactly what he or she means”).
root construct, combinable with different leadership Likewise, there is behavior–intent conflation for rela-
styles (Avolio & Gardner, 2005). Subsequently, there tional transparency (e.g., “makes decisions based on
was a period of conceptual consensus around the his or her core values”). To assert that a leader really
definition and measure proposed by Walumbwa says what they mean or whether they act upon core
et al. (2008) that established authentic leadership values, respondents compare observed behaviors
as a distinct leadership style focusing on moral with attributions of presumed intentions and values.
self-concordance (Lemoine et al., 2019). Finally, in There is also behavior–execution conflation for
recent years scholars have started challenging both measuring what Walumbwa et al. (2008) referred
the conceptual consensus and the very idea of to as an “internalized moral perspective” (e.g.,
authentic leadership (Alvesson & Einola, 2019). “listens carefully to different points of view before
Reviewing the dominant conceptualization. coming to conclusions”) and as “balanced proc-
Walumbwa et al. (2008) put forward a definition and essing” (e.g., “accurately describes how others view
measure of authentic leadership that has dominated his or her capabilities”). Listening and describing
the field of research and had received nearly 4,500 refer to the content of behaviors, whereas carefulness
Google Scholar citations as of February 2022. Draw- and accuracy refer to evaluating whether this content
ing on Luthans and Avolio (2003), Walumbwa et al. was executed well. Building on the conceptualization
(2008: 94) specified that authentic leadership “draws of Walumbwa et al. (2008), an alternative measure
342 Academy of Management Annals January

of authentic leadership is the Authentic Leadership of authentic leadership. By contrast, Hoch et al. (2018)
Inventory (ALI). The ALI seeks to address psychomet- and Banks et al. (2016) questioned incremental valid-
ric shortcomings of the ALQ (Neider & Schriesheim, ity vis-a-vis other leadership constructs. More funda-
2011), yet draws on the same conceptualization as the mentally, Alvesson and Einola (2019) criticized the
ALQ. Thus, the ALI is prone to the same operational conceptual underpinnings of authentic leadership
shortcomings as the ALQ. Indeed, items such as “My and warned about using authentic leadership and
leader expresses his/her ideas and thoughts clearly to other leadership constructs that have excessively posi-
others” and “My leader uses his/her core beliefs tive labels. In a letter exchange, proponents and critics
to make decisions” (Neider & Schriesheim, 2011: discussed whether authentic leadership is useful
1149) exemplify conflation via evaluative qualifiers (Gardner, Karam, Alvesson, & Einola, 2021). Our
(“clearly”) and unobservable elements requiring an review does not take a stance on whether it is good
inference (“core beliefs”). or bad to be authentic, but we have identified that
Reviewing and assessing alternative conceptual- authentic leadership maps on to multiple aspects (i.e.,
izations according to the four aspects framework. descriptive and evaluative) of leadership. Such map-
There are a variety of alternative conceptualiza- ping constitutes valence-based conflation and causal
tions preceding the work of Walumbwa et al. indeterminacy, which calls into question the existing
(2008). For instance, Bass and Steidlmeier (1999) evidence base. In this way, our review critiques the
popularized the idea of authenticity in leadership current authentic leadership concept because we
research when distinguishing “authentic trans- uncover conflation as the root cause of limitations in
formational” from “pseudo-transformational” lead- authentic leadership without suggesting that leaders
ership based on whether the forms of leadership are should be inauthentic (as done, e.g., by Pfeffer, 2015).
“ethical” (i.e., authentic) or “unethical” (i.e., pseudo) Summary. Authentic leadership research helped
(for a critique, see Price, 2003). Subsequently, Luthans to raise scholarly awareness about ethical concerns
and Avolio (2003: 243) defined authentic leadership in leadership and led to much empirical research on
as “a process that draws from both positive psycholog- the topic affirming its positive consequences. We
ical capacities and a highly developed organizational find it plausible that authentic leadership causes
context, which results in both greater self-awareness positive outcomes because authenticity can improve
and self-regulated positive behaviors on the part credibility, which in turn can help in wielding social
of leaders and associates, fostering positive self- influence (see also Lind & Sitkin, 2015). However,
development.” Additional work on authentic leader- our review recognizes that the dominant conceptuali-
ship built on this conceptualization (e.g., Avolio, zation of authentic leadership research is limited by
Gardner, Walumbwa, Luthans, & May, 2004; Gard- behavior–execution conflation. Its measurement tool
ner, Avolio, Luthans, May, & Walumbwa, 2005). is, in addition, limited by behavior–intent conflation.
This initial definition is also positively valenced and These forms of valence-based conflation raise con-
maps on to the content and quality of executing cerns about the validity of the evidence base. Decon-
behaviors (“process that draws from [ … ] positive flating the authentic leadership construct will be
psychological capacities”) as well as effects (“results an important step toward producing more rigorous
in [ … ] greater self-awareness and self-regulated pos- research and knowledge concerning authentic leader-
itive behaviors”). There are other approaches that we ship in the future.
do not discuss here because they have regarded
authentic leadership as a root construct rather than a Empowering Leadership
leadership style, and hence are beyond our scope.
Evolution of the concept. The idea of empow-
Examples are an embodiment perspective on
authentic leadership (Ladkin & Taylor, 2010), life ering leadership dates back to classical works on par-
story– and narration-based approaches (Shamir & ticipative management (Likert, 1961) and supportive
Eilam, 2005; Sparrowe, 2005), and popular science leadership (Bowers & Seashore, 1966). It is attractive
applications (George, 2003). because of its promise to simultaneously foster
Reviewing previous critiques. Recent reviews employee performance and satisfaction. However,
and critiques have focused primarily on work in the in contrast to authentic leadership, there is not one
tradition of Walumbwa et al. (2008) because it pro- consensual definition and measure of empowering
duced the largest and most coherent body of research. leadership. Articles in our sample draw on four
A review by Lemoine et al. (2019) criticized some parts competing conceptualizations (see Ahearne, Mathieu,
of authentic leadership research, such as certain ques- & Rapp, 2005; Amundsen & Martinsen, 2014; Arnold,
tionnaire items, but largely affirmed the effectiveness Arad, Rhoades, & Drasgow, 2000; Kirkman &
2023 Fischer and Sitkin 343

Rosen, 1999). The work of Ahearne et al. (2005) has support, and development support with intent to
received slightly more traction than the other three promote their experience of self-reliance, motiva-
approaches, yet all four approaches continue to be tion, and capability to work autonomously within
regularly used. Notwithstanding conceptual differ- the boundaries of overall organizational goals and
ences, all four approaches share a critical commonal- strategies.” The definition spans behavioral content
ity: that empowering leaders are supportive and (e.g., “power sharing”), intentions (“with the intent
increase followers’ autonomy in a nonburdening way. to”), and quality of execution to produce certain
Reviewing the dominant conceptualization. The effects (“promote their experience [ … ] to work auton-
most influential conceptualization of empowering omously”). The consequence is behavior–intent and
leadership is that by Ahearne et al. (2005). They behavior–execution conflation.
defined and measured empowering leadership A number of studies have used the measure of exter-
behaviors around four dimensions: enhancing the nal team leader behaviors based on a definition pro-
meaningfulness of work, fostering participation in posed by Kirkman and Rosen (1999) to operationalize
decision-making, expressing confidence in high per- empowering leadership (e.g., Chen, Sharma, Edinger,
formance, and providing autonomy from bureaucratic Shapiro, & Farh, 2011; Harris, Li, Boswell, Zhang, &
constraints. The effectiveness of this version of empow- Xie, 2014; Li, Chiaburu, Kirkman, & Xie, 2013). Specif-
ering leadership has been associated with increased ically, Kirkman and Rosen (1999: 59) defined team
creativity (Zhang & Bartol, 2010) or decreased cynicism empowerment as having four dimensions: “potency,
(Lorinkova & Perry, 2017). In addition, meta-analytic meaningfulness, autonomy, and impact,” and external
evidence has shown that various versions of empower- team leader behaviors are seen as antecedent to such
ing leadership improve task performance and organiza- team empowerment.
tional citizenship behavior (Lee, Willis, & Tian, 2018; Summary. Despite variation between the four dif-
see also Cheong et al., 2019). ferent versions of empowering leadership, they all rest
Mapping the dominant conceptualization ac- on leader support and follower autonomy, and are all
cording to the four aspects framework to assess supposed to cause positive outcomes (Cheong et al.,
valence-based conflation. Ahearne et al.’s (2005) 2019). While the idea of empowering leadership is
four dimensions of leadership behaviors have an laudable, all four conceptualizations are limited by
inherently positive valence and map onto not only valence-based conflation and span behavioral content,
the content aspect of leadership behaviors but also as well as good execution. Furthermore, one version
good execution and effects. For instance, enhanced of empowering leadership is defined such that it pre-
meaning and removed constraints are effects of leader- supposes empowering intentions (Amundsen &
ship behaviors. Thus, only behaviors that are either Martinsen, 2014) and two versions presuppose
executed well or intended to enhance meaning empowering effects (Ahearne et al., 2005; Arnold
or remove constraints at work qualify as empowering et al., 2000). There is a need for nonconflated con-
and meet definitional standards. Taken together, ceptualizations in the future.
Ahearne et al.’s (2005) definition conflates behavioral
content with effects and quality of execution or inten- Ethical Leadership
tions (see also Table 3).
Reviewing and assessing alternative conceptual- Conceptual evolution. Ethical leadership emerged
izations according to the four aspects framework. as a research topic in response to an alarming num-
Arnold et al. (2000: 1240) defined empowering lead- ber of corporate scandals (Trevi~ no, Hartman, &
ership as “behaviors whereby power is shared with Brown, 2000), and has sparked broad research
subordinates and that raise their level of intrinsic interest (Lemoine et al., 2019). Empirical ethical
motivation.” Power sharing specifies behavioral and leadership research is characterized by strong con-
decisional content (“shared”), while raising intrinsic sensus around the definition and measure proposed
motivation is an effect (“raise their level”) and pre- by Brown et al. (2005), which rests on interpersonal
supposes that the behaviors are executed well. Thus, behaviors and compliance with normative stand-
the definition maps onto multiple behavioral aspects ards. In our sample, all but one empirical article (i.e.,
and suffers from behavior–execution and behavior– Kalshoven, Den Hartog, & De Hoogh, 2011) built on
effect conflation. the Brown et al. (2005) conceptualization. Neverthe-
Amundsen and Martinsen (2014: 489) defined less, there is conceptual controversy about whether
empowering leadership as “the process of influenc- ethical leadership is a leader behavioral or follower
ing subordinates through power sharing, motivation perceptual phenomenon (Den Hartog, 2015;
344 Academy of Management Annals January

Eisenbeiss, 2012), and Banks et al. (2021a) have different views, behavioral and evaluative views exist
taken the position that ethical leadership as pro- in parallel. On one hand, most empirical research has
posed by Brown et al. (2005) conflates both views. conceptualized ethical leadership in terms of leader-
Reviewing the dominant conceptualization. ship behaviors (e.g., Ng & Feldman, 2015). On the
Brown et al. (2005: 120) defined ethical leadership as other hand, several scholars have explicitly theorized
“the demonstration of normatively appropriate con- and tested an evaluation-focused approach to ethical
duct through personal actions and interpersonal rela- leadership (Jordan, Brown, Trevi~ no, & Finkelstein,
tionships, and the promotion of such conduct to 2013; Stouten, van Dijke, Mayer, De Cremer, &
followers through two-way communication, reinforce- Euwema, 2013), which was also reflected in a recent
ment, and decision-making.” They operationalized review (Den Hartog, 2015).
this definition with the 10-item Ethical Leadership Reviewing previous critiques. One advantage of
Scale (ELS). This measure has been used, for instance, perceptual or evaluative views of ethical leadership
by Mayer, Aquino, Greenbaum, and Kuenzi (2012) is that they can be applied to very different contexts
who studied both antecedents and consequences of and cultures. Eisenbeiss (2012), by contrast, criti-
ethical leadership. In addition, a meta-analysis (Ng & cized such views as relativistic. Indeed, defining
Feldman, 2015) and a narrative review (Lemoine et al., ethical leadership by follower perceptions causes
2019) both reported positive effects of ethical leader- ambiguity of underlying behaviors because even
ship, but their findings did not take into account objectively abhorrent behaviors might be classified
valence-based conflation and causal indeterminacy. as ethical depending on the norms of followers.
Mapping the dominant conceptualization ac- Banks et al. (2021a) identified conflation between
cording to the four aspects framework to assess behaviors and evaluations as a fundamental impedi-
valence-based conflation. Brown et al.’s (2005) defi- ment for producing rigorous evidence about ethical
nition maps onto behavioral content (“conduct”), leadership. These critiques raise the concern that,
and it is ambiguous whether it covers not only ethi- although meta-analyses and reviews have affirmed
cal intentions but also effects (“promotion of such positive effects of ethical leadership (Lemoine et al.,
conduct”). To the extent that a behavior is more 2019; Ng & Feldman, 2015), such findings might be
likely to be classified as promotion of normatively an artifact of conflation. In other words, the percep-
appropriate conduct if the behavior was intended to tual legacy of ethical leadership makes the concept
or has the effect of producing such follower-level prone to valence-based conflation because it shifts
behavior, the definition of ethical leadership is lim- the focus from the behavior to how the perceived
ited by behavior–intent and behavior–effect confla- behavior is evaluated—and such evaluation can
tion. In the case of behavior–effect conflation, there unfold along presumed intentions, quality of execu-
is also behavior–execution conflation because achiev- tion, and produced effects.
ing ethical goals presupposes good execution of the Summary. Ethical leadership might be a solution
behaviors. Thus, the risk of all three types of valence- to moral wrongdoing in organizations, and indeed
based conflation is present for the definition of ethical scholars have found that it is linked to many positive
leadership (see also Table 3). The three types of confla- outcomes (Ng & Feldman, 2015). It is unclear, how-
tion are present in a clearer form in the questionnaire ever, whether the existing evidence base about ethi-
measure (ELS). Some items specify the behavior by its cal leadership is valid because, as a valenced
underlying intentions (e.g., “Has the best interests of concept, ethical leadership is prone to conflating
employees in mind”), other items by high-quality exe- leader behaviors with follower evaluations (Banks
cution (e.g., “Makes fair and balanced decisions”). et al., 2021a). To produce a more rigorous evidence
Additional items also implicate presumed effects base, future research needs to build on deconflated
(e.g., “Can be trusted”). Thus, past research has offered concepts that distinguish behavioral descriptions
empirical evidence based on a conflated measure. and evaluations.
Reviewing and assessing alternative conceptual-
izations according to the four aspects framework.
Servant Leadership
The work of Trevi~ no et al. (2000) was a predecessor
of Brown et al. (2005) and studied the development Conceptual evolution. The concept of servant
of a reputation for ethical leadership. Thus, the same leadership dates back to the early work of Greenleaf in
authors conceptualized ethical leadership once in the 1970s. Despite the decades-long history, there is a
behavioral (i.e., conduct-based) and once in evalua- lack of consensus in studying servant leadership (Eva,
tive (i.e., reputation-based) terms. Reflecting these Robin, Sendjaya, van Dierendonck, & Liden, 2019).
2023 Fischer and Sitkin 345

Among the various definitions and measures of the execution. Multiaspect conflation also manifests in
construct, our review uncovered two particularly questionnaire items. For instance, “My leader gives
influential ones: those of Ehrhart (2004) and Liden, me the freedom to handle difficult situations in the
Wayne, Zhao, and Henderson (2008). Notwithstand- way that I feel is best” (Liden et al., 2015: 256) specifies
ing conceptual differences, both views center around both behavioral content and quality of execution;
patterns of interpersonal behaviors, enacted with a “My leader makes my career development a priority”
consequentialist focus on creating valued outcomes (Liden et al., 2015: 256) refers to both what a leader
for followers and other stakeholders (Lemoine et al., does and the idea that the behavior is executed well;
2019). and “I would seek help from my leader if I had a per-
Reviewing the dominant conceptualization. The sonal problem” (Liden et al., 2015: 256) is an effect of
conceptualization and measurement tool proposed leadership behaviors. By implication, Liden and col-
by Liden et al. (2008) sparked most research interest. leagues’ (Liden et al., 2015; Liden et al., 2008) concep-
Liden et al. (2008: 162) argued that servant leaders tualization and measurement of servant leadership is
focus “on developing employees to their fullest limited by valence-based conflation.
potential in the areas of task effectiveness, community Assessing an alternative conceptualization ac-
stewardship, self-motivation, and future leadership cording to the four aspects framework. Ehrhart
capabilities.” Instead of a more formal definition, (2004) developed another variant of the servant lead-
they put forward an operational conceptualization of ership construct that rests on recognizing moral
servant leadership along seven dimensions that responsibility beyond contributing to organizational
are also reflected in the questionnaire measure (i.e., performance and considering various stakeholders,
SL-28; four items for each of the seven dimensions, including subordinates. Without offering a formal
hence 28 items overall): conceptual skills, empower- definition, Ehrhart (2004) developed a 14-item mea-
ment, helping subordinates grow and succeed, put- sure that reflects the additional focus on moral
ting subordinates first, behaving ethically, emotional responsibility but remains limited by valence-based
healing, and creating value for the community. There conflation because the measure spans all four as-
is also a shortened 7-item questionnaire (SL-7), which pects of leadership: intention (e.g., “makes the per-
includes one item of the SL-28 for each of its seven sonal development of department employees a
dimensions (Liden et al., 2015). Scholars have reported priority”), content (e.g., “my department manager
empirical support for the effectiveness of the servant does what she or he promises to do”), quality of exe-
leadership construct (e.g., Chiniara & Bentein, 2016; cution (e.g., “balances concern for day-to-day details
Lemoine & Blum, 2021; Liden, Wayne, Liao, & Meuser, with projections for the future”), and effect (e.g.,
2014). However, the presence of valence-based confla- “creates a sense of community among department
tion raises questions about the foundation for claims employees”) (Ehrhart, 2004: 93). Conflation thus
about its validity, which we assess next. calls into question the validity of empirical support
Mapping the dominant conceptualization ac- for the effectiveness of Ehrhart’s (2004) servant lead-
cording to the four aspects-framework to assess ership concept (e.g., for studies by Neubert, Hunter,
valence-based conflation. The seven dimensions of & Tolentino, 2016; Neubert et al., 2008; Walumbwa,
servant leadership span multiple aspects of leader- Hartnell, & Oke, 2010).
ship behaviors (see also definitions of these dimen- Summary. There seems to be broad empirical sup-
sions in Liden et al., 2008: 162). The conceptual port for the positive consequences of servant leader-
skills dimension is not behavioral but is indirectly ship (Lemoine et al., 2019), yet both dominant
linked to the quality of executing behaviors. The conceptualizations are limited by valence-based
dimensions empowerment, helping subordinates conflation. It should be noted that one article in our
grow, and emotional healing span behavioral con- sample pointed toward a promising alternative
tent, quality of execution, and effects because approach. Sun (2013) conceptually examined the
empowerment, subordinate growth, and health are identity-based underpinnings of servant leadership,
leadership outcomes. Moreover, putting subordi- arguing that a multiattribute servant identity influen-
nates first refers to both prosocial intention and ces cognitive dispositions that, depending on the
behavioral content, and behaving ethically mixes the organizational context, influence behavioral disposi-
content of behaviors with laudable intentions (if deon- tions and servant leadership. Such work can help in
tological) and the quality of execution plus effects (if dissecting servant leadership into its nonvalenced
consequentialist). Furthermore, creating value for the antecedents and constituent components, foresha-
community specifies effects and presupposes good dowing a deconflated research approach.
346 Academy of Management Annals January

Charismatic Leadership rhetoric. Building on earlier work by House (1977),


Shamir, House, and Arthur (1993) argued that charis-
Conceptual evolution. Charismatic leadership is
matic leaders influence followers by affecting their
the oldest concept that we review and predates the
self-concept. Shamir, Arthur, and House (1994)
modern study of leadership, for example in Max
applied this view in suggesting that charismatic lead-
Weber’s (1922/2002) work characterizing charisma as
ers use seven rhetorical techniques (e.g., “More refer-
one source of authority. Building on, yet also deviating
ences to the collective and to collective identity, and
from Weber’s sociological premises, House (1977) sug-
fewer references to individual self-interest” [Shamir
gested a psychological theory of charismatic leadership
et al., 1994: 29]) to influence followers’ self-concept.
that popularized the notion of charisma among man-
Following this approach, Bligh, Kohles, and Meindl
agement scholars. Over the last two decades, two (2004) and Baur et al. (2016) used computer software
research streams have been particularly popular: (a) to content-analyze speeches. Offering an example of
studying charismatic leadership as a subconcept of promising ways to move forward more rigorously,
transformational leadership, and (b) studying charis- Baur et al. (2016) identified regularly cooccurring
matic leadership as leader rhetoric. During the period patterns of rhetorical techniques as configurations,
covered by this review, the most common approach has which they examined to assess the differential suc-
been to study charisma as part of transformational lead- cess of these configurations in wielding influence.
ership. However, in recent years rhetorical approaches This rhetoric-based research approach does not suf-
to charismatic leadership have gained popularity. fer from valence-based conflation because it focuses
Reviewing the dominant conceptualization. on “what” a leader says—not necessarily in terms of
Scholars have commonly conceptualized charismatic substance but in terms of rhetoric and words—and
leadership as “idealized influence” and “inspirational hence does not mix the description of behaviors with
motivation,” which are to have two dimensions of the evaluation of underlying intentions, quality of
transformational leadership and are typically measured execution, or effects.
with the MLQ (Antonakis et al., 2016). This conceptual- More recently, Antonakis, d’Adda, Weber, and
ization is identical to the conceptualization of these two Zehnder (2021) and Jacquart and Antonakis (2015)
dimensions of transformational leadership that we measured charisma with different rhetorical tactics,
reviewed previously. A meta-analysis (Banks, Enge- such as anecdotes, contrasts, expressing moral con-
mann, Williams, Gooty, McCauley, & Medaugh, 2017) victions, metaphors, or setting ambitious goals, and
showed that charismatic leadership predicts positive defined charisma as “values-based, symbolic, and
outcomes, including individual and group perfor- emotion-laden leader signaling” (Antonakis et al.,
mance. However, Banks et al. (2017) noted that the 2016: 304). Their approach avoids conflation, as
MLQ-based primary studies have suffered from concep- does other recent empirical research outside our
tual and empirical limitations—notably lack of clarity sample (Meslec, Curseu, Fodor, & Kenda, 2020; Tur,
as to whether the measures capture behaviors, and Harstad, & Antonakis, 2021). Whereas the seven rhe-
endogeneity bias—that put in doubt the validity of torical strategies by Shamir et al. (1994) share a focus
the findings. on evoking a more collective self-concept, the set of
Mapping the dominant conceptualization ac- rhetorical tactics by Antonakis and colleagues is
cording to the four aspects framework to assess broader and more eclectic because the tactics are so
valence-based conflation. The conceptualization of general that some of them are not specific to leader-
charisma as idealized influence and inspirational ship (e.g., use of anecdotes, contrasts, metaphors).
motivation, like transformational leadership research, Still, the approach is promising for future research
is characterized by valence-based conflation (see also because it overcomes all three types of valence-based
our previous mapping of transformational leadership conflation by capturing only what leaders say, and
and the two charisma dimensions). This conflation avoiding confusion with intentions, execution, or
manifests by spanning multiple aspects of leadership effects.
behaviors (for further review-based critiques, see There are also alternative conceptualizations of
Antonakis et al., 2016; Banks et al., 2017; Van Knip- charismatic leadership that are neither MLQ- nor
penberg & Sitkin, 2013). rhetoric-based, but these have received less academic
Reviewing and assessing alternative conceptual- traction. Thus, we do not include them in this review.
izations according to the four aspects framework. Examples are charismatic leadership as part of the
The most popular alternative approach to studying so-called CIP model distinguishing charismatic, ideo-
charismatic leadership is based on the leader’s logical, and pragmatic leadership (Lovelace, Neely,
2023 Fischer and Sitkin 347

Allen, & Hunter, 2019; Strange & Mumford, 2005); transactional) relationships and appeal to followers’
charisma based on media descriptions (Flynn & Staw, values and ideals (i.e., transformational) but also has
2004); and studying antecedents of charisma in terms to ensure that followers contribute to the organiza-
of affective events (Walter & Bruch, 2009). tion’s strategy (Antonakis & House, 2014). In this
Reviewing previous critiques. A set of articles way, instrumental leadership has been suggested as
criticized the MLQ-based dominant conceptualiza- extending transactional-transformational leadership
tion of charismatic leadership and its valence-based to reflect more than the so-called “full range of lead-
conflation, including the above-discussed works by ership” (Bass & Bass, 2008: 879)—to reflect the
Yukl (1999) and Van Knippenberg and Sitkin (2013). “fuller full range of leadership” (Antonakis & House,
In addition, Antonakis et al. (2016: 306) spoke of the 2014: 746). Please note that we review only the ver-
“MLQ problem,” arguing that the MLQ used for mea- sion proposed by Antonakis and House (2014)
suring charisma does not measure an independent because it is the only conceptualization of instru-
variable but an outcome. Banks et al. (2017) lamented mental leadership in our sample.
similar conceptual and empirical limitations. These Reviewing the dominant conceptualization.
critiques cast doubt on the validity of research on Antonakis and House (2014: 749) defined instrumen-
charismatic leadership in the tradition of the MLQ tal leadership as “the application of leader expert
and serve as foundations for the scholarly shift toward knowledge on monitoring of the environment and of
studying charisma rhetorically. performance, and the implementation of strategic and
Summary. The notion of charismatic leadership tactical solutions”; they built on this definition to
has a long history, yet the heterogeneity of approaches observe that, “strategically, leaders monitor the exter-
makes it impossible to characterize charismatic lead- nal environment and identify strategies and goals.
ership research in a monolithic way. Still, there is a From a follower work facilitation point of view, lead-
clear trend in our sample. Over the entire review ers provide direction and resources, monitor perfor-
period, studying charismatic leadership as a subcon- mance and provide feedback.” Following current
cept of transformational leadership has been the most terminology, this definition refers to management
popular approach; however, this focus has shifted in more than to leadership (Ashford & Sitkin, 2019;
recent years toward studying charisma as leader rhe- Kotter, 1990), yet this explanation captures the four
toric. Scholars are apparently beginning to heed the dimensions of instrumental leadership: environ-
calls of Van Knippenberg and Sitkin (2013) and Anto- mental monitoring, strategy formulation, path– goal
nakis et al. (2016) to avoid confounding behaviors facilitation, and outcome monitoring. Environmen-
with effects, and the study of charismatic leadership tal monitoring rests on “knowing the strengths and
is a positive exemplar for studying leadership styles weaknesses of the organization and identifying
in a deconflated manner. A remaining challenge for opportunities,” while strategy formulation refers to
future research, though, is to ensure theoretical inte- “leader actions focused on developing policies, goals,
gration. We will outline later how a configurational and objectives to support the strategic vision and mis-
approach can be helpful for addressing this concern. sion” (Antonakis & House, 2014: 750). In addition,
path–goal facilitation refers to “leader behaviors tar-
Instrumental Leadership geted towards giving direction, support, and
resources, removing obstacles for goal attainment
Conceptual evolution. Instrumental leadership is and providing path–goal clarifications,” and outcome
a recently formulated leadership style (Antonakis & monitoring “entails leader provision of performance-
House, 2014) that has its roots in path–goal theory, enhancing feedback useful for goal attainment”
another classic in the study of leadership (House, (Antonakis & House, 2014: 750).
1971; House & Mitchell, 1975). Path–goal theory, Mapping the dominant conceptualization ac-
which holds that effective leaders “engage in behav- cording to the four aspects framework to assess
iors that complement subordinates’ environments valence-based conflation. The four dimensions of
and abilities in a manner that compensates for defi- instrumental leadership specify what a leader does,
ciencies and is instrumental to subordinate satisfac- and hence they capture the content of leadership
tion and individual and work unit performance” behaviors. In addition, the dimensions of path–goal
(House, 1996: 323), has received meta-analytic sup- facilitation (e.g., “removes obstacles”) and outcome
port based on 120 studies (Wofford & Liska, 1993). monitoring (e.g., useful and “performance-enhancing
The idea of instrumental leadership is that a leader feedback”) span quality of execution and effects. Envi-
not only has to engage in fair exchange (i.e., ronmental monitoring and strategy formulation also
348 Academy of Management Annals January

span quality of execution insofar as behaviors are Reviewing the dominant conceptualization. Tep-
more likely to be classified as belonging to the per (2000: 178) defined abusive supervision as
two dimensions when the behaviors are helpful for “subordinates’ perceptions of the extent to which
fostering the respective goals (e.g., “identifying oppor- supervisors engage in the sustained display of hos-
tunities” or “support[ing] the strategic vision”). tile verbal and nonverbal behaviors, excluding phys-
The questionnaire for measuring instrumental ical contact.” He specified that this definition
leadership reflects the mix of multiple behavioral characterizes abusive supervision as a subjective
aspects too (see Antonakis & House, 2014: 753). For assessment. The same individual could view a
instance, the items “ensures that his/her vision is supervisor’s behavior as abusive in one context and
understood in specific terms” (strategy formulation as non-abusive in another context, and two subordi-
and implementation) and “removes obstacles to my nates could differ in their evaluations of the same
goal attainment” (path–goal facilitation) conflate supervisor’s behavior (Tepper, 2000: 178). Several
behaviors with effects because follower understanding corresponding questionnaire items, by contrast, are
and removed obstacles are outcomes of leadership formulated as descriptions of behaviors (e.g., “Breaks
behaviors. Likewise, “provides me with constructive promises he/she makes” [Tepper, 2000: 190]). The
feedback about my mistakes” (outcome monitoring) response scales used in abusive supervision ques-
conflates behavioral content (“feedback”) with good tionnaire items are also behavioral (e.g., response
execution (“constructive”) because the item not only option “He/she uses this behavior very often with
describes that a leader gives feedback but asks for eval- me” [Tepper, 2000: 189]).
uating whether the feedback was given well to be con- Generally, there is ambiguity as to whether abu-
structive. In addition, “understands the constraints of sive supervision rests on leadership behaviors and is
our organization” (environmental monitoring) does a leadership style or is a subjective evaluation of fol-
not specify a behavior but a high-quality sense of lowers. On one hand, the definition proposed by
judgment. Tepper (2000) refers to abusive supervision as a sub-
jective evaluation, and Tepper et al. (2017: 126) reaf-
Summary. The idea of instrumental leadership is firmed that the abusive supervision construct does
appealing: leaders also have to foster contributions not “capture supervisors’ behavior in an objective
to the organization’s strategy (Antonakis & House, sense.” On the other hand, a recent systematic
2014). Whereas we concur with this general idea and review has shown that abusive supervision is still
find it a valuable addition to leadership research, we predominantly referred to and studied as a set of
raise concerns about the inherent positive valence of leader behaviors, including in work by Tepper him-
instrumental leadership for fostering organizational self (see Fischer et al., 2021). For instance, scholars
goals and its resulting conceptual conflation. have described the behavioral domain of abusive
supervision (Yu, Duffy, & Tepper, 2018), and meta-
Abusive Supervision analyses have included the hypothesized conse-
quences of abusive supervision (Mackey et al., 2017;
Conceptual evolution. The concept of abusive Zhang & Liao, 2015). Thus, notwithstanding the con-
supervision is the most influential approach to ceptual debates about whether abusive supervision
studying negative forms of leadership and has is behavioral or evaluative, we include abusive
helped draw attention to the downsides of leader- supervision in this review because it has been com-
ship influence. Abusive supervision research is monly used as a construct capturing patterns of lead-
characterized by a strong consensus around the ership behaviors that are evaluated and believed as
definition and measure provided by Tepper (2000). fostering undesirable consequences, and, in this
Furthermore, all empirical articles on abusive way, has been characterized as a leadership style.
supervision in our sample have built on Tepper’s Mapping the dominant conceptualization accord-
conceptualization. Integrative reviews and meta- ing to the four aspects framework to assess valence-
analyses (Mackey, Frieder, Brees, & Martinko, 2017; based conflation. In mapping abusive supervision
Zhang, Liu, Xu, Yang, & Bednall, 2019) have sug- according to the four aspects framework, we find
gested that abusive supervision leads to negative that the abusive supervision construct conflates the
workplace outcomes, yet a recent review identified descriptive with evaluative aspects of leadership
conflation in abusive supervision research, raising behaviors. Some items explicitly commingle behav-
concerns about the validity of this conclusion iors with underlying intentions (“Blames me to save
(Fischer et al., 2021). himself/herself embarrassment” [Tepper, 2000: 190]),
2023 Fischer and Sitkin 349

constituting behavior–intent conflation. Other items of negatively valenced leadership style research.
conflate descriptions with negative evaluations of Einarsen et al. (2007) put forward a construct defini-
leadership behavior more ambiguously (“Gives me tion of destructive leadership centered around the
the silent treatment”: [Tepper, 2000: 189]). Because violation of legitimate organizational interests.
precisely the same behavior can be interpreted differ- Destructive leadership is also often invoked as an
ently depending upon who exhibits it (e.g., if my boss umbrella term for the “dark side of leadership” in
does not interact much with me because they have general, spanning distinct concepts such as petty
confidence in my work versus because they do not tyranny, toxic leadership, or simply bad leadership
care about how I am doing), this negative evaluation (Tierney & Tepper, 2007).
can rest on ascribed meaning that implies different Reviewing dominant conceptualizations.
presumed intent (i.e., inadvertently silent behavior) Einarsen et al. (2007: 208) conceptualized destructive
or not (i.e., deliberately silent behavior). In the former leadership as “the systematic and repeated behaviour
case, there is behavior–intent conflation. In the latter by a leader, supervisor or manager that violates the
case, the behavior is inadvertently poorly executed legitimate interest of the organisation by undermining
for maintaining proper relationships at work, such and/or sabotaging the organisation’s goals, tasks,
that there is behavior–execution conflation. Fischer resources, and effectiveness and/or the motivation,
et al. (2021) outlined for each item of Tepper’s (2000) well-being or job satisfaction of subordinates.” Both
questionnaire how it conflates behaviors and evalu- meta-analytic evidence (Schyns & Schilling, 2013)
ations and offers a more comprehensive assess- and an integrative narrative review (Krasikova et al.,
ment of conflation. Here, we simply record the 2013) have provided empirical support for the nega-
existence of valence-based conflation. tive effects of destructive leadership. The validity of
Reviewing previous critiques. Abusive supervi- such evidence depends, however, on the validity of its
sion research has found broad support. Integrative conceptualization and measurement (Antonakis, 2017).
reviews (Tepper, 2007; Tepper et al., 2017) and Mapping the dominant conceptualization ac-
meta-analyses (Mackey et al., 2017; Zhang et al., cording to the four aspects framework to assess
2019) have reported consistently negative effects valence-based conflation. The Einarsen et al. (2007)
of abusive supervision. However, confidence in the definition spans multiple aspects of leadership, which
meta-analytic evidence rests on the rigor of the under- can be seen at the example of “sabotaging” (Einarsen
lying primary studies (Antonakis, 2017). Fischer et al. et al., 2007: 208). Deliberate sabotaging refers to a neg-
(2021) have shown that conflation, as well as low base ative intent, and nondeliberate “sabotaging” refers to
rates of reported abusive supervision, undermine con- very poor execution of leader behaviors, which
fidence in the statistical inferences drawn by meta- implies either behavior–intent or behavior–execution
analytic evidence and underlying primary studies conflation. In addition, “violat[ing] the legitimate
alike (e.g., Harris, Kacmar, & Zivnuska, 2007; Tepper, interests of an organization” (Einarsen et al., 2007:
Henle, Lambert, Giacalone, & Duffy, 2008). 208) refers to a negative effect. Thus, the definition
Summary. Abusive supervision is an important exhibits all three types of valence-based conflation. In
topic of research because such a form of leadership their integrative review, Krasikova et al. (2013: 3)
might engender very negative effects for people and explicitly built on Einarsen et al. (2007) and defined
organizations. However, our review identifies that, destructive leadership “as volitional behavior by a
despite the clean original definition offered, its leader that can harm or intends to harm a leader’s
implementation since that time has resulted in the organization and/or followers.” This definition con-
current conceptualizations of abusive supervision flates behaviors with the evaluation of underlying
that suffer from behavior–intent, behavior–execu- intentions (“volitional,” “intends”), as well as poten-
tion, and behavior–effect conflation, which casts tially negative effects (“harm”).
doubt on the validity of the evidence base. Deconflat- Reviewing and assessing alternative conceptual-
ing the abusive supervision construct would be a izations according to the four aspects framework.
positive step in enhancing confidence in this leader- Other scholars have examined single facets of
ship style construct. destructive leadership or taken distinct approaches
to conceptualizing it, which could be an avenue for
deconflating destructive leadership. For instance,
Destructive Leadership
Padilla, Hogan, and Kaiser (2007) took a systemic
Conceptual evolution. Destructive leadership is approach and regarded destructive leadership as
another commonly studied and well-cited stream the interplay of leaders, susceptible followers, and
350 Academy of Management Annals January

conducive environments (see also Thoroughgood, work of Padilla et al. (2007) that distinguished the toxic
Padilla, Hunter, & Tate, 2012). Ferris, Zinko, Brouer, triangle—destructive leaders, susceptible followers,
Buckley, and Harvey (2007), in turn, focused on one and conducive environments—might be helpful in
subfacet of destructive leadership, such as strategic this regard for keeping leader behaviors, follower eval-
bullying, which still mixes behaviors with intentions. uations, and contextual features conceptually apart.
Others have taken a more exploratory approach
(Shaw, Erickson, & Harvey, 2011) or treated destruc-
SYNTHESIZING THE SYSTEMATIC REVIEW OF
tive leadership as rooted in abusive supervision (Col-
LEADERSHIP STYLES
lins & Jackson, 2015). These different views share the
research logic of spanning behavioral content as well In this section, we first briefly summarize general
as one further aspect of leadership, be it underlying insights of our review and take stock of valence-
intentions or behavioral effects. based conflation. Then, we outline how valence-
Summary. Whereas abusive supervision centers on based conflation leads to amalgamation, construct
the experience of abuse by a follower, destructive lead- redundancy, and causal indeterminacy. Figure 1
ership adds a complementary focus on violating orga- summarizes the key findings of our review.
nizational interests (Einarsen et al., 2007). There is
meta-analytic evidence for the negative effects of these
General Insights of the Review
forms of leadership (Schyns & Schilling, 2013), but the
validity of such research is unclear due to valence- The systematic review has shown that each of the
based conflation. Thus, there is a need for research 10 leadership styles sheds light on an important sub-
based on deconflated constructs that focus on one topic of leadership, ranging from ethics (i.e., authen-
aspect of leadership behaviors only. The conceptual tic and ethical leadership), follower involvement

FIGURE 1
Overview of the Common Research Logic, Conceptual Weaknesses, and Way Forward
for Leadership Style Research

Amalgamation
Otherwise conceptually
unrelated behaviors are bundled
in the same leadership style
when they share valences

Evaluative logic:
Studying what good/bad Bundling behaviors Valence-based conflation Construct redundancy
leaders do into a leadership style Classifying behaviors not only Many constructs with similar
Flaw of studying what leaders based on content but also valence populate—and overlap
in part based on
do in a charged sense valences (good or bad intentions, on—the limited construct space
goodness or badness
quality of execution, or effects) of leadership styles

Causal indeterminacy
Obfuscating whether content,
good or bad intentions, execution,
or effects relate to the outcome
(tautologies are one type, or sub-
set, of causal indeterminacy)

Overcoming valence-based conflation and causal indeterminacy: Studying leadership styles in a de-conflated and configurational maner
1. De-conflation along the four aspects-framework: Distinguish between intended leadership styles, displayed leadership style, as well as their realized effects
2. Configuration: Keeping single components conceptually apart yet specifying a causal model how these components relate to each other

Notes: The solid lines indicate that one concern necessarily leads to the other (e.g., valence-based conflation necessarily leads to causal inde-
terminacy), and the dashed lines indicate that one concern increases the risk/probability of the other concern (e.g., valence-based conflation
promotes the risk of amalgamation).]
2023 Fischer and Sitkin 351

and support (i.e., empowering and servant leader- fair and balanced decisions” (Brown et al., 2005: 125)
ship leadership), higher values (i.e., charismatic and serve as illustration. Leadership training that pre-
transformational leadership), relations- versus task- scribes fair and balanced decision-making would be
orientation (i.e., consideration and initiating structure perfectly aligned with compiled evidence on ethical
leadership), organizational strategy (i.e., instrumental leadership (Bedi, Alpaslan, & Green, 2016; Lemoine
leadership), follower abuse (i.e., abusive supervision), et al., 2019). However, the prescription is vague to the
to violating legitimate interests of the organization point of being difficult to teach and frustrating to learn.
(i.e., destructive leadership). From the perspective of a training participant, the
Leadership style research has produced a seem- question arises as to what one is supposed to do when
ingly consistent evidence base, suggesting that posi- trying to follow this prescription of ethical leadership.
tive leadership styles lead to positive outcomes (i.e., For instance, is it fair or unfair, and is it balanced or
do-good logic) and that negative leadership styles unbalanced (a) to fire an employee after months of
lead to negative outcomes (i.e., don’t-do-bad logic). underperformance, (b) to grant an employee a pay raise
Such findings have drawn an uplifting image of above the levels of colleagues to entice them to stay, or
leadership style research and leadership practice, (c) to give an employee with mental health issues pref-
affirming scientific progress and negating potential erential treatment in terms of performance standards?
tensions between doing good and doing well (e.g., Answers to these questions are likely to vary depend-
Avolio et al., 2009; Lemoine et al., 2019). As our sys- ing on national and organizational culture, as well as
tematic review shows, these findings might be an the business environment, but they are also likely be
artifact of valence-based conflation and causal inde- contested within many organizations. Training partici-
terminacy. Supporting evidence of the do-good and pants might be looking for guidance on which deci-
don’t-do-bad logic of leadership rests on shaky con- sions are fair and balanced and on how to make such
ceptual grounds, which puts into question the valid- decisions, instead of getting the generic (and barely
ity of the findings. Overall, our review has identified actionable) advice that decisions should be fair and
a set of problems that characterize all 10 reviewed balanced. Moving on from our general insights, we
leadership styles and that stem from valenced con- next summarize the major sources of conflation identi-
ceptualizations that conflate the description of leader- fied in our review, and the consequences that result.
ship behavior with the evaluation of leadership styles
(see also Tables 1, 2, 3). Thus, our review supports
arguments of those who have expressed skepticism Taking Stock of Valence-Based Conflation
about the strong claims in extant leadership research
(e.g., Alvesson & Einola, 2019; Antonakis et al., 2010; We find that valence-based conflation is nearly a
Pfeffer, 2015; Van Knippenberg & Sitkin, 2013). defining feature of how leadership styles have been
This general insight also raises questions about studied to date. All 10 reviewed leadership styles
leadership development and training that is based span both behavioral content and quality of execution
on findings of leadership style research limited by and are hence limited by behavior–execution confla-
the concerns identified in this manuscript. We advo- tion. In addition, two leadership styles explicitly
cate evidence-based management education (see include intentions in both their conceptualization
also DeRue, Sitkin, & Podolny, 2011b; Rousseau, and measurement (ethical and servant leadership),
Manning, & Denyer, 2008), and warn against man- and intentions are part of measurement items of other
agement education that rests on evidence whose styles as well (i.e., authentic and servant leadership
validity is limited by conflated conceptualizations and abusive supervision).
and causal indeterminacy. Such leadership develop- The conceptualizations of eight leadership styles
ment and training has two types of problems: a (all except consideration-initiating structure leader-
flawed evidence base, and ambiguous prescriptions. ship and authentic leadership) conflate behaviors
First, the content of such training might be wrong with their effects. This has already been documented
because, as outlined above, the do-good and the as a serious problem by Van Knippenberg and Sitkin
don’t-do-bad logic of leadership style research is not (2013) for the cases of charismatic and transactional-
supported by sound evidence. Second, the prescrip- transformational leadership. Our review shows that
tions of such training are likely to be ambiguous behavior–effect conflation limits the confidence
because valence-based conflation obfuscates, or even in most constructs of leadership style research and
obstructs, what exactly has to be done to lead in a is a much more widespread problem than com-
certain style. Ethical leadership and the item “makes monly presumed. Next, we examine problematic
352 Academy of Management Annals January

consequences of valence-based conflation for leader- decision-making independent of any relational ties.
ship style research. Authentic leadership involves not only relational
transparency as a set of interpersonal behaviors but
Taking Stock of Amalgamation also self-awareness and an internalized moral perspec-
tive, which are in the realm of information processing
Our review has shown that when researchers con- and manifest in judgment and decision-making. Using
flate the content of behavior with the valence of Mintzberg’s (1973) terminology, such leadership styles
the behavior (i.e., evaluations of presumed underly- mix interpersonal with informational and decisional
ing intentions, quality of execution, or behavioral roles. Or, in more recent terminology, they amalgam-
effects), they are likely to bundle together multiple ate interpersonal behaviors, which are in the realm
conceptually distinct behaviors in their constructs. of leadership, with decision-making, which is in
Thus, valence-based conflation weakens the concep- the realm of management (Ashford & Sitkin, 2019;
tual rationale for classifying leadership behaviors Bedeian & Hunt, 2006).
into leadership styles. By implication, otherwise
conceptually unrelated behaviors are amalgamated
Taking Stock of Construct Redundancy
in the same leadership style when they share valen-
ces. For instance, the same tough feedback can be Valence-based conflation is also a source of con-
part of ethical leadership or abusive supervision struct redundancy in leadership research. Construct
depending on the presumed intention (e.g., ethical redundancy is the result of two or more constructs
leadership if the tough feedback is judged as norma- overlapping in their construct domain and hence
tively appropriate or helpful to the recipient, and having limited incremental variance in explaining
abusive if judged as rude or unnecessarily harmful to outcomes relative to each other. In a meta-analysis,
the recipient). Tepper (2000) noted that such evalua- Hoch et al. (2018) demonstrated that authentic and
tions are subjective and vary across people who ethical leadership have limited incremental validity
make judgments. Likewise, vastly different behav- relative to transformational leadership, indicating
iors can be part of ethical or destructive leadership, redundancy between these constructs. Banks et al.
as long as presumed underlying intentions or behav- (2018) argued that such limited incremental validity is
ioral effects are positive (for ethical) or negative (for due to overlap in the behaviors described by various
destructive), thus meeting the spirit of the respective leadership styles.
leadership style. The consequence is that valenced Extending these insights, we find that leadership
leadership styles become conceptual hodgepodges of styles overlap in not only their descriptions but
all kinds of otherwise unrelated behaviors that evalua- also their evaluations of leadership behaviors. For
tors judge as sharing positive or negative valences. instance, empowering and servant leadership over-
Because this is a flawed justification for the creation of lap in attention to the quality of executing behavior
a coherent “style,” amalgamation weakens conceptual supportive of followers and in achieving corre-
precision and coherence and undermines the ability sponding effects (Ahearne et al., 2005; Liden et al.,
to specify the shared mechanisms by which they have 2008). In addition, construct redundancy can also
predictable effects (Hirsch & Levin, 1999; Van Knip- result from nonoverlapping yet diametrically
penberg & Sitkin, 2013). Although amalgamation has opposed construct spaces when two constructs share
been lamented previously (Carton, 2022; Van Knip- the same focus for assessing valence (e.g., normative
penberg & Sitkin, 2013), in this review we have identi- compliance) but have opposed directionality (i.e.,
fied valence-based conflation as a continuing root positive versus negative). That is, two variables may
problem in the leadership style literature. appear to be distinct, but are to a large degree simply
One benefit of the kind of synthetic review per- the opposite poles of the same continuum. For
formed here is that it also helps to clarify that amalgam- instance, in a given context, a behavior cannot be a
ation is particularly pronounced in leadership style manifestation of ethical and destructive leadership
conceptualizations that are not delimited to inter- at the same time because they are conceptualized as
personal behaviors but also involve a leader’s judg- mutually exclusive: the former rests on a leader’s
ment, information processing, and decision-making. legitimate (i.e., normatively appropriate) conduct
For instance, in addition to interpersonal path–goal (Brown et al., 2005) and the latter on a leader violat-
clarifying behaviors, instrumental leadership in- ing legitimate interests of the organization (Einarsen
cludes strategy formulation and implementation et al., 2007). By definition, an increase in destructive
(Antonakis & House, 2014), which is in the realm of leadership entails a decrease in ethical leadership.
2023 Fischer and Sitkin 353

Empirically, both forms of leadership should strongly Our analyses show that conceptualizations of positive
correlate, even though with a negative sign. By impli- and negative leadership styles not only tap into leader
cation, a strong correlation between two predicting behaviors but also represent valenced evaluations of
variables leads to limited incremental validity because a leader’s character, competence, and success. Thus,
one construct has limited explanatory power vis-a-vis construct redundancy in leadership styles goes beyond
the other construct. overlapping descriptions of behaviors.
In general, the eight reviewed positive leadership
style research streams and the two negative leader- Taking Stock of Causal Indeterminacy
ship style research streams share an evaluative
approach even though the directionality of valence How valence-based conflation leads to causal
changes. This overlap might be explained by three indeterminacy. Most problematically, valence-based
types of root causes: commonalities in evaluations of conflation creates causal indeterminacy in the rela-
a good or bad character, high or low leadership com- tionship between a leadership style (the independent
petency, and high or low leadership success. These variable) and outcomes (the dependent variable),
three potential root causes are intended to capture because conflation obfuscates which aspects of a
the general tendencies of a person with regard to the leadership style (i.e., behavioral content, good or bad
three evaluative aspects such that good or bad char- intentions, good or bad execution, or good or bad
acter is associated with a higher base rate of good effects) relate to the focal outcome (e.g., good or
or bad intentions, high or low competency with a bad follower performance or satisfaction). In other
higher base rate of good or poor quality of execution, words, by definition, valence implies a conflation of
and high or low success with a higher base rate descriptions of leadership with evaluations of lead-
of positive or negative effects. Particularly in the ership, rendering it impossible to decipher which
absence of rich situation-specific information about aspect of leadership is causally linked to the out-
a leader’s intentions, quality of execution, or effects, come of interest. Even worse, evaluations of leader-
raters of leadership styles have to draw on more gen- ship, notably evaluations of its behavioral effects, are
eral evaluations of the leader’s character, compe- causally more proximate to the focal outcome than
tency, or general success to evaluate the leader’s the description of behavioral content (see Figure 2).
style. That is, in a specific situation, judging a leader Causal proximity between evaluations of leader-
as having a good or bad character might be the source ship and other outcomes of leadership (e.g., follower
of attributing good or bad intentions, judging a leader satisfaction) makes findings related to evaluations
as high or low in competence might be the source of and focal outcomes prone to being affected by shared
evaluating a person as good or bad in terms of execu- antecedents (i.e., common cause) and mutual influ-
tion, and judging a leader’s success might be the ence (i.e., reverse or bidirectional causality). For
source of inferring good or bad behavioral effects. instance, evaluating a leader as having a good charac-
Consequently, assessments of leadership styles ter might be a common cause of a follower attributing
characterized by behavior–intent conflation (e.g., ethi- positive intentions to the leader and being satisfied
cal leadership) might largely tap into evaluations of a at work; a follower evaluating a leader as highly com-
leader’s character, which is in line with Trevi~ no et al. petent might be a common cause of evaluating the
(2000), who put center-stage a leader’s reputation for a leader’s behavior as well-executed and being satis-
moral character. Leadership styles characterized by fied at work; and a follower evaluating a leader as
behavior–effect conflation (i.e., all except authen- generally successful might be a common cause of
tic and consideration-initiating structure leader- evaluating a leader as transformative (i.e., behavioral
ship) might largely be assessed by evaluating a effect) and being satisfied at work. However, a com-
leader’s success, which is what has often occurred mon cause between evaluations and outcomes of
in research demonstrating that information about pre- leadership implies that there can be a style–outcome
vious success increases ratings of positive leader- link irrespective of the displayed leadership
ship styles (e.g., Fischer & Dietz, 2020; Lord et al., behaviors. Put differently, cooccurrence of good or
1978). Given that all leadership style research con- bad evaluations and good or bad outcomes of leader-
flates behaviors with quality of execution, the defini- ship does not imply causal effects of leadership behav-
tions of leadership styles routinely conflate the iors, which is illustrated by our previous example
quantity with the quality of displaying a certain lead- of the manager displaying hard-hitting behaviors for
ership behavior, and not only describe what leaders changing customer service procedures. Employees
generally do but also evaluate a leader’s competence. being receptive versus resistant to the change might
354 Academy of Management Annals January

FIGURE 2
Valence-Based Conflation, Causal Indeterminacy, and the Pattern of Flawed Results
Antecedents of Descriptive aspect of Evaluative aspects of Focal outcome of
leadership leadership leadership leadership

Attribution of good or bad


? intentions

Traits of the leader and


follower
?

Good or bad focal outcome


Behavioral content (e.g., ? Evaluation of good or bad ? (e.g., follower
words) execution of behaviors performance or satisfaction,
org. performance)

Features of the context ?


(including leader–
follower relationship) ?
Evaluation of good or bad
effects

Positive or negative leadership style (e.g., empowering and


transformational leadership or destructive leadership)

Causal proximity to Causal proximity to


antecedents of leadership outcomes of leadership

The erroneous logic of studying positive/negative leadership styles as if they were unconflated descriptions of leadership behaviors

Do-good logic Positive leadership style Good focal outcome

Don’t-do-bad logic Negative leadership style Bad focal outcome

Notes: The further a box is positioned to the right, the more causally proximate the corresponding concept is to the focal outcome and the
more the concept is an outcome itself. Causal proximity between two constructs makes it likely that they share antecedents or mutually influ-
ence each other, which renders the relationship between the two constructs causally indeterminate. Further, if a focal outcome is defined by or
identical to an evaluative aspect (e.g., empowering or transformational effects partially defining a focal outcome such as follower empowerment
or motivation), the relationship between a valenced leadership style and the focal outcome is tautological. A tautology is a subset of causal
indeterminacy.

account both for the focal outcome (e.g., employee per- positive outcomes, and negative leadership with
formance, satisfaction) and for judgments of a leader negative outcomes.
as having transformative or abusive effects. In Likewise, good intentions, well-executed behav-
addition, related to the well-known performance iors, and proximate behavioral effects might not
cue effect, success in implementing a new proce- only cause a focal outcome such as follower job satis-
dure informs judgments of transformative and abu- faction but also be the result of that outcome (i.e.,
sive effects, which constitutes reverse causality. bidirectional causality). That is, people use out-
Thus, irrespective of the causal effect of the hard- comes such as follower job satisfaction as cues for
hitting behavior itself, shared antecedents and evaluating or inferring the character (related to
mutual influence can make for a causally indeter- intentions), competence (related to quality of execu-
minate cooccurrence of positive leadership with tion), and success (related to effectiveness) of leaders
2023 Fischer and Sitkin 355

(cf. Epitropaki, Foti, & Hansbrough, 2020). This means, Notwithstanding the type of causal indeterminacy,
however, that outcomes of leadership often inform our analysis shows that valence-based conflation ob-
what are supposed to be a priori and independent scures the exact causes of leadership style– outcome
evaluations of leadership styles. Thus, valence-based links and that the empirical support for the do-good
conflation obfuscates the causes of leadership style– and the don’t-do-bad logic rests on shaky conceptual
outcome relationships. grounds. Cooccurrence of positive evaluations and
This obfuscation follows a pattern. As our system- positive outcomes of leadership, as well as negative
atic review has shown, positive leadership styles have evaluations and negative outcomes of leadership, tend
been found to “predict” positive outcomes and nega- to coincide for many other reasons than the causal
tive leadership styles have been found to “predict” effects of leadership.
negative outcomes. Although positive evaluations of Why causal indeterminacy is a genuinely con-
leadership have been found to coincide with positive ceptual problem. In empirical analyses, causal inde-
outcomes of leadership, and negative evaluations of terminacy might appear as an endogeneity problem.
leadership have been found to coincide with negative In statistical terms, common causes are called omit-
outcomes of leadership, this finding should not be ted variables while bidirectional causality is called
mistaken to necessarily imply a causal relationship. a simultaneity problem—two classical sources of
The do-good logic, as well as the don’t-do-bad logic, endogeneity (Hill, Johnson, Greco, O’Boyle, & Wal-
can be an artifact of causal indeterminacy manifested ter, 2020). In recent years, endogeneity has become a
through common causes of or bidirectional causal- topic of increasing interest to leadership scholars,
ity between evaluative aspects of leadership and and more robust statistical analysis tools have been
leadership outcomes. For instance, leader behav- suggested to address endogeneity problems and
ior tends to be evaluated more positively if the ensure proper estimation results (e.g., Antonakis
leader is, for whatever reasons, associated with et al., 2010; Sajons, 2020). We endorse the trend
positive outcomes (akin to the well-known perfor- toward stronger research methods. Still, we warn
mance cue effect [Gioia & Sims, 1985; Lord et al., that research methods alone are insufficient for over-
1978]), which illustrates bidirectionality by switch- coming the problem of causal indeterminacy
ing presumed causal direction. In addition, positive because the root cause is valence-based conflation.
ratings of leadership styles and positive outcomes For instance, if the quality of executing leadership
could both be caused by a positive organizational behaviors is defined in terms of observed leadership
or economic environment, which illustrates third- performance, it is simultaneously a behavior and the
variable causation. outcome of that behavior (Campbell, 2012; Campbell
In the most extreme case of causal indeterminacy, & Wiernik, 2015). In addition, attributions or evalua-
the leadership style–outcome relationship is a tau- tions of presumably underlying intentions of leader-
tology because the outcome is defined by the style ship commonly arise after observing the leadership
(cf. Popper, 1959). For instance, it would be tautolog- behaviors and outcomes, and are hence outcomes
ical to theorize that empowering leadership be- too (Martinko et al., 2007; Meindl et al., 1985).
haviors empower followers if behaviors that are Linking attributed intentions, quality of execution,
classified as empowering leadership are defined in or effects with an outcome such as follower perfor-
terms of empowering effects. Likewise, it would be mance means linking a more proximal outcome (i.e.,
tautological to theorize that destructive leadership the evaluative aspects of leadership) with a more dis-
reduces follower performance if destructive leader- tal outcome (i.e., follower performance). Empirical
ship behaviors are defined by having negative analyses appear biased due to endogeneity resulting
effects on followers. In these cases, by definition from omitted variables or simultaneity problems.
“good” behaviors (e.g., empowering) lead to “good” However, the origin of such endogeneity is not sta-
outcomes (i.e., empowerment) and “bad” behaviors tistical modeling but valence-based conflation,
(e.g., destructive) lead to “bad” outcomes (i.e., poor because a predictor variable—the content of leader-
performance). Generally, behavior–effect conflation ship behaviors—is mixed up with proximal outcomes
makes leadership style research prone to tautologies such as the attribution of intentions or observed effects
because effects are built into the conceptualization in a positively or negatively valenced leadership style.
of leadership styles (e.g., empowerment, destruc- The statistical model testing a leadership style–
tion), and these effects can overlap with the out- outcome link is only misspecified because leadership
come of interest (e.g., follower performance or styles are conflated and are not valid representations
satisfaction). of leadership behaviors.
356 Academy of Management Annals January

DISSECTING THE LOGIC OF 70 YEARS OF Four Sublogics of Studying What Good and Bad
LEADERSHIP STYLE RESEARCH Leaders Do
In this section, we first document that valence- In the systematic review, we have shown that lead-
based conflation is not an idiosyncratic weakness ership styles capture different variants of goodness
of only the reviewed leadership styles—it also or badness in a leader’s pattern of behaviors. We now
occurs in four general sublogics of leadership style outline four sublogics under which all 10 leadership
that we have identified in leadership research, style research streams can be subsumed—the study
which have emerged sequentially over the last 70 of what (a) effective leaders do, (b) value-driven lead-
years, and currently coexist. Despite their differ- ers do, (c) moral leaders do, and (d) immoral leaders
ences, these four sublogics share a valenced con- do. We argue that these four sublogics offer only var-
ceptualization that specifies what good or bad iations of valence-based conflation in studying what
leaders do instead of specifying what leaders do. good leaders or bad leaders (but not both) do. We also
Second, we show that valence-based conflation is indicate what it would take to overcome these weak-
reflected in the general conceptualization of the nesses and really study what leaders (irrespective of
notion of leadership style and reinforced by the whether they are good or bad) do by distinguishing
commonly used practice of measuring leadership behaviors from their evaluations. In addition, we
with questionnaires. Throughout this section, we sketch alternative approaches that have already been
point to alternatives for better research practice, dissected and have a long research tradition yet have
before expanding on these ideas in the subsequent disappeared from the mainstream leadership research
section. landscape. Table 4 summarizes the key points.

TABLE 4
Varieties and Commonalities in Sublogics of Leadership Style Research
Valence-Based
Sublogics of Leadership Encompassed Valenced Principles of Conceptual Weaknesses Conceptual Remedies
Style Research Leadership Styles the Sublogicsa of the Sublogics for the Sublogics

Effective or functional Consideration and Leadership that meets Specifying leadership Dissecting the
leadership styles initiating structure needs to foster behavior by its behavioral construct
(positive, leadership higher performance function for from its performance
performance-related Empowering or related improving (-related) effects
valence) leadership organizationally performance or
Instrumental relevant variables related variables
leadership (e.g., job performance
Transactional or job satisfaction)
leadership
Value-based leadership Charismatic leadership Value-based leadership Specifying leadership Dissecting the
styles (positive, Transformational that engages or behavior by its behavioral construct
influence-based leadership inspires followers inspirational impact from its inspirational
valence) beyond expectations on followers impact
or self-interest
Moral leadership styles Authentic leadership Leadership based on the Specifying leadership Dissecting the
(positive normative Ethical leadership “right” values behavior by its moral behavioral construct
valence) Servant leadership (deontological or righteousness from its moral
virtue-based) or evaluation
causing “good”
outcomes (i.e.,
consequentialist)
Immoral leadership Abusive supervision Leading based on the Specifying leadership Dissecting the
styles (negative Destructive leadership “wrong” values or behavior by its moral behavioral construct
normative valence) causing “bad” despicability from its moral
outcomes evaluation

a
For simplicity, we specify here the sublogic based on their positive or negative effects (e.g., follower performance or well-being).
Commonly, leadership styles have also been specified by good or bad intentions or quality of execution for reaching the effect (Table 3
offers an overview). However, whether it is behavior–intent, behavior–execution, or behavior–effect conflation, the basic logic remains the
same: specifying behaviors based on their goodness/badness.
2023 Fischer and Sitkin 357

Effective or functional leadership styles refer to These approaches have also been labeled “neo-
positive contributions to organizational perfor- charismatic” or “new leadership” because they
mance. Consideration and initiating structure complement older effectiveness-based versions of
leadership dates back to research conducted more leadership (Antonakis & House, 2013). For instance,
than 70 years ago and distinguishes relations- and House (1977) formulated charismatic leadership as a
task-oriented behaviors (see, e.g., Stogdill, 1950). complement to path–goal theory. Bass’s (1985) con-
These behaviors are supposed to make employees cept of transformational leadership was intended to
more effective by meeting their interpersonal and “augment” transactional leadership. Whereas it is
task-related needs, respectively. Such a needs-based valuable to examine leaders’ appeals to values and
conceptualization is even more explicit in path–goal their impact on followers and their performance,
theory (House, 1971, 1996) and in conceptualiza- most previous research went further by defining
tions of transactional leadership (Bass, 1985), such value-based behaviors in terms of their capacity to
that leaders have to meet the needs of followers to inspire followers to go beyond self-interest (see
make them perform better. Instrumental leadership, a Table 4). This problem has been already identified
spin-off of path–goal theory, expands this logic from by Van Knippenberg and Sitkin (2013) and other
researchers, who called for separating behaviors
the follower level to the organizational level (Antona-
from their effects. Antonakis et al. (2016) outlined
kis & House, 2014). Similarly, empowering leadership
such an approach for charismatic leadership. This
is about offering conditions that make followers more
approach has been increasingly used in primary stud-
effective in autonomously solving problems and tak-
ies (e.g., Antonakis et al., 2021; Meslec et al., 2020), as
ing initiative for making positive contributions to the
well as meta-analytic research (Ernst et al., 2021), and
organization (Ahearne et al., 2005). allows for studying what leaders do in terms of value-
Fleishman et al. (1991) synthesized such work based appeals, without specifying how well they do
into a framework of functional leadership behaviors it. We believe that such an approach is more fruitful
where behaviors contribute to the functioning of the than conflation-based research practice.
organization, its subunits, or individual members. Moral leadership styles refer to unambiguously
Yukl (2012) offered a related framework, asserting positive normative achievements or compliance.
that effective leadership behaviors produce positive Since the 1990s, scholars have become increasingly
effects. The problem with these frameworks and wary about the potentially immoral consequences
their underlying primary research is that specifica- of value-based forms of leadership because leaders
tions of effective or functional behaviors have a might appeal to socially undesirable values or appeal
built-in positivity because behaviors are classified to desirable values but abuse these values for unde-
based not only on their content but also on their qual- sirable ends. Thus, scholars have started to distin-
ity of execution or even their positive performance guish moral from immoral value-based approaches.
effects (see Table 4). While functional specifications of For instance, by splitting charisma into a positive
leadership behaviors can be prone to conflation, func- socialized and a negative personalized form (Howell
tional theorizing is still very valuable for synthesizing & Avolio, 1992), or by splitting transformational
existing research or hypothesizing about yet-to-be leadership in a positive authentic and a negative
examined causal relationships (see, e.g., Morgeson, pseudo-transformational form (Bass & Steidlmeier,
DeRue, & Karam, 2010). 1999), scholars have suggested a variety of moral
Value-driven leadership styles refer to norma- leadership styles now including authentic, ethical,
tive achievements or compliance. Since the 1970s, and servant leadership (Lemoine et al., 2019). The
scholars have drawn attention to leadership behav- basis for judging the morality of behaviors varies
iors that are driven by values rather than by func- across these leadership styles but shares the problem
tional or effectiveness considerations. Whereas of conflating distinct aspects of leadership because
functional approaches are about meeting people’s behaviors are specified by their moral righteousness
needs to unleash their known performance potential, (see Table 4). Thus, moral leadership is an extreme
value-based approaches such as charismatic and case of valence and studying what “good” lead-
transformational leadership are about achieving ers do.
“performance beyond expectations” (Bass, 1985). Immoral leadership styles refer to unambigu-
According to these theories, leaders who appeal to fol- ously negative normative achievements or (non-)
lowers’ values can motivate and influence followers compliance. Study of immoral forms of leadership
to go beyond self-interest (Bass, 1985; House, 1977). took off in the 2000s with the formulation of abusive
358 Academy of Management Annals January

supervision (Tepper, 2000). The research logic is managerial work relate to performance is lacking,
similar to that of moral leadership, with inverse and, thus, has only occasionally been referenced in
directionality of valence. Behaviors are classified as the mainstream leadership research reviewed here.
immoral when they rest on reprehensible values or The Lewin- and Mintzberg-type research offers
have harmful consequences (see Table 4). Even alternatives to studying what good or bad leaders do
though studying immoral forms of leadership might in a valenced manner, but do suffer from their own
appear to be dramatically different from the study of limitations (see, e.g., Willmott, 1984). Still, these
positive forms of leadership, it exhibits the same forms of research have illustrated that there are via-
valenced approaches and would similarly benefit ble alternatives to the valenced approaches of study-
from distinguishing behaviors from their evaluation. ing good or bad leadership that dominate today.
Alternatives to valenced leadership style con-
structs. There are two alternative approaches that are The Overarching Logic of Studying What Good
descriptive of what leaders do: (a) leadership style or Bad Leaders Do
research in the tradition of Kurt Lewin (1948/1997)
We have systematically reviewed 10 leadership
and further developed by Vroom & Jago (1988), and
styles and traced the evolution of four dominant sub-
(b) research on managerial work, initially pioneered
logics of leadership style research. However, we
by Sune Carlson and further developed by Henry
have not yet assessed the leadership style concept in
Mintzberg (see the overview by Tengblad, 2003).
general and how its typical measurement produces
First, Lewin, Lippitt, and White (1939) pioneered
spurious results. We do this next to offer a compre-
the study of leadership and experimentally manipu-
hensive assessment of leadership style research.
lated autocratic, democratic, and laissez-faire leader- The common leadership style definition speci-
ship displayed by adult leaders in a children’s camp. fies what good or bad leaders do. Although the
They found that a democratic style caused less ag- concept of a leadership style is commonly invoked,
gression than an autocratic style, while the laissez- it is rarely explicitly defined (see, e.g., Antonakis &
faire leadership was in between. Their finding of Day, 2017; Yukl & Gardner, 2019), with an important
positive consequences of democratic leadership exception. The Bass Handbook of Leadership pos-
sparked interest among psychologists and manage- ited that “leadership and management styles are
ment scholars. For instance, Tannenbaum and alternative ways that leaders and managers pattern
Schmidt’s (1973) popular science article on how to their interactive behavior with those they influence”
choose a leadership pattern advanced the work by (Bass & Bass, 2008: 73–74). This definition is a fair
Lewin et al. (1939) by treating boss-centered (i.e., auto- representation of past and current usage of the term
cratic) and subordinate-centered (i.e., democratic) leadership style, both in common parlance (Mer-
leadership as a continuum instead of a dichotomy. riam-Webster, 2020) and in scholarly work (e.g.,
Vroom and Jago (1988) built upon Lewin’s work by Anderson & Sun, 2017).
suggesting contingency factors for when it is more The Bass and Bass (2008) definition provides a
effective to be directive versus consultative. While springboard for assessing the overarching logic of
Kurt Lewin’s (1948/1997) research was driven by a leadership styles more generally. The definition has
search for social betterment and for democratizing four components: leadership, management, patterns
society, his research on leadership styles was not lim- of interactive behavior, and influence. The first, sec-
ited by valence-based conflation because he systemati- ond, and third components distinguish leadership
cally described and manipulated behavioral content from management, interactive behavior from unilat-
(see Table 1 in Lewin et al., 1939). eral decision-making, and patterns of behavior from
Second, in the 1950s Sune Carlson’s pioneering single concrete behaviors. The fourth component
research (cf. Tengblad, 2003) formed the basis for distinguishes influence from authority-based direc-
Mintzberg’s (1973) well-known descriptive studies tion. Especially relevant for our purpose is the notion
of managerial work. Key findings are that there is a of influence. By including the term “with those they
high pace of work, brevity of managerial activities, influence” (Bass & Bass, 2008: 74), this conceptuali-
high action orientation, and preference for informal, zation of leadership styles presupposes influence.
lateral communication (Mintzberg, 2009). Research Due to the notion of influence, studies of leadership
on managerial work has avoided the problems iden- styles have investigated what good or bad leaders do
tified in the 10 reviewed leadership styles. However, instead of what leaders in general do. Pfeffer (1977)
quantitative assessment of how different forms of lamented ambiguity in the concept of leadership
2023 Fischer and Sitkin 359

because the concept is so big that it is hard to distin- leads to causal indeterminacy in a systematic man-
guish leadership from social influence. We agree that ner. A good evaluation—whether of a leader’s inten-
there is no leadership—and, for that matter, no lead- tion, quality of execution, or effects—is a positive
ership style—without social influence, but warn that proximal outcome, and a negative evaluation is a
mixing a behavior with the effect or intent to influ- negative proximal outcome. Thus, when raters give
ence somebody else makes the concept prone to high scores on a positive leadership style measure or
conflation. a high score on a negative leadership style measure,
Leadership styles include all kinds of influence they indicate their approval or disapproval of the
toward achieving objectives that are relevant for a leader. It is no surprise that such measures correlate
follower, team, or organization in the broadest sense with other, more distal outcomes such as follower
(Fleishman et al., 1991; Yukl, 2012). Effective performance or satisfaction. The do-good logic and
or functional leadership styles refer to influencing the don’t-do-bad logic of leadership can exist as
performance-related variables, value-based leader- artifacts of conflated measurement, irrespective of
ship styles refer to influencing followers beyond actual causal relationships. Please note that this con-
self-interest, and (im-)moral leadership styles refer cern exists independent of raters and their relation-
to influencing followers or organizations in a mor- ship with the leader or their cognitive biases because
ally laudable or reprehensible way. In some cases, the problem exists at the level of the item and hence
leadership styles additionally refer to the intention precedes problems of responding to such items.
to achieve these different types of objectives (e.g.,
ethical or servant leadership). Thus, generalizing
THE WAY FORWARD: DECONFLATING
beyond our systematic review, and generalizing
THROUGH A CONFIGURATIONAL APPROACH
even beyond the four identified sublogics, we can
say that leadership styles have been conceptualized Based on our review of research on leadership
in the past and are commonly studied today as pat- styles, we suggest that future work should apply a
terns of informal and interactive behaviors that foster deconflated and configurational approach. Such an
certain follower-, team-, or organization-related approach would retain the benefits of a broad and
objectives. This conceptualization is problematic multifaceted conception of a leadership style while
because style-specific influence blends what leaders avoiding the conflation-induced traps of current
do with the study of how good versus bad leaders approaches. In this section, we first offer an outline
exercise influence based on good versus bad motives for studying leadership styles in a deconflated man-
or achieve good versus bad outcomes; valence-based ner. Then, we sketch the key pillars of a deconflated,
conflation and causal indeterminacy is the inevitable configurational view. Subsequently, we discuss im-
consequence. plications for the empirical study of leadership
The common leadership style measures specify styles. We conclude with a general note on the future
what good or bad leaders do. In addition to the of leadership style research.
problematic general conceptualization of leadership
styles, there is a problematic approach to measure- Studying Leadership Styles in a Deconflated
ment that segregates good behaviors from bad behav- Manner
iors, and good intentions from bad intentions—and
then focuses on evaluations of a leader’s style that Beyond naïve holism. Our review uncovered an
samples on the pre-valenced dependent variable. In emergent view in leadership research that warns
addition, these studies have mixed behaviors with about the perils of overly broad constructs. For
attributions of intention and behavioral valance. instance, calls have been made for using more
Typical leadership style research has measured refined concepts in the study of authentic (Alvesson
leadership behaviors with follower ratings through a & Einola, 2019), ethical (Banks et al., 2021a), and
set of questionnaire items (next to our systematic transformational leadership (Van Knippenberg & Sit-
review, see also Hunter, Bedell-Avers, & Mumford, kin, 2013), as well as abusive supervision (Fischer
2007). In line with the conflated conceptualization et al., 2021). Our review has shown that the valenced
of leadership styles, however, these items formulate specification of leadership styles has fostered amal-
what good or bad leaders do and conflate the descrip- gamation and construct redundancy, and, even more
tion of behaviors (i.e., what a leader does) with the problematically, has rendered leadership style re-
evaluation of those behaviors (i.e., whether they are search causally indeterminate. Hirsch and Levin
good or bad). As outlined previously, conflation (1999) proposed that broad umbrella concepts are
360 Academy of Management Annals January

only appropriate in early stages of a research pro- Studying Leadership Styles in a


gram but that maturing fields need to use increas- Configurational Manner
ingly refined concepts. The field of leadership
Beyond scattered reductionism. While deconflat-
research is now well established, and it is high time
ing leadership styles is necessary, we do not call for
to go beyond valence-based conflation in the study
its extreme opposite, which would be atomizing
of leadership styles.
leadership research into scattered reductionism with
Distinguishing intended and displayed leader- minuscule constructs. Because the fundamental idea
ship styles, and their realized effects. To overcome of a style is not about individual elements but
conflation, we follow a previous conceptual devel- instead concerns recognizable patterns among those
opment in strategy research. Whereas initially schol- elements, advocating atomization would reject the
ars used strategy as a broad umbrella term that spans very idea of leadership styles as “alternative ways
planning, the activity of strategizing, and effective that leaders [ … ] pattern their interactive behavior
strategic positioning, Mintzberg and Waters (1985) with those they influence” (Bass & Bass, 2008;
distinguished intended, emergent, and realized strat- 73–74). As “patterns” of behaviors, leadership styles
egies (for a recent update, see Mirabeau, Maguire, & are irreducible to any single behavior. Thus, concep-
Hardy, 2018). Analogous to Mintzberg and Waters’s tually integrating scattered behaviors is necessary.
(1985) approach to deconflating the umbrella term Katz and Kahn (1978), Van Knippenberg and Sitkin
“strategy,” we suggest deconflating the umbrella (2013), and Lind and Sitkin (2015) have already
term “leadership style.” That is, we suggest distin- explicitly called for studying leadership styles con-
guishing intended leadership styles from displayed figurationally, and the work of Kipnis, Schmidt, and
(i.e., emergent) leadership styles, and from their real- Wilkinson (1980) and Yukl and Falbe (1990) on influ-
ized effects. ence tactics demonstrated that leadership behaviors
We suggest the following definitions of leadership can be classified in conceptually meaningful bundles
styles. (i.e., configurations), and that these meaningful bun-
dles—and not isolated behaviors—have causal effects
 Intended leadership styles refer to the patterns of on outcomes of interest. In this way, a configurational
informal and interactive behaviors a leader wants approach can serve as a middle ground that avoids
to display (e.g., empowering intentions). conflation yet ensures conceptual integration.
 Displayed leadership styles refer to displayed pat- Whereas leadership has not yet been studied config-
terns of interactive behaviors (e.g., content of exe- urationally, future scholarship can build on the con-
cuting empowering behaviors). figurational tradition in adjacent fields.
 Realized effects of leadership styles refer to the The configurational tradition in management
achieved results that are believed to be causally research. In human resource management and orga-
driven by the pattern of leadership behaviors nization theory, configurational research has a long
(e.g., empowering effects). tradition that can be tapped and applied to leader-
ship (for overviews, see Delery & Doty, 1996; Misan-
The three leadership style concepts share a focus gyi, Greckhamer, Furnari, Fiss, Crilly, & Aguilera,
on patterns of interactive behaviors that are sup- 2017). For instance, Ichniowski, Shaw, and Pre-
posed to foster certain goals. However, the three nnushi (1997) used a configurational lens to demon-
refined leadership style concepts avoid conflation strate empirically that specific bundles of distinct
of multiple aspects of leadership. For instance, human resource management practices can lead to
intended leadership styles focus only on one aspect increases in productivity, but only when imple-
of leadership: underlying intentions. Likewise, dis- mented jointly. More generally, Miller (1986, 1996,
played leadership styles focus only on behavioral 2018) formulated configurational research as a dis-
content, and realized effects of leadership styles tinct paradigm, and recently a configurational para-
focus only on effects of leadership. Generally, all digm for testing and theorizing has been further
three types of leadership styles are worth studying in developed (e.g., Fiss, 2007; Fiss, Marx, & Cambr e,
their own right. In addition, studying the interrela- 2013). Very recent applications to job demands (Ong
tionships among these types of styles can be illumi- & Johnson, 2021) and corporate venture innovations
nating. However, to avoid causal indeterminacy, it is (Waldkirch, Kammerlander, & Wiedeler, 2021) have
critical to keep the three types of leadership style suggested that this paradigm is beginning to flourish
conceptually and empirically apart. among organizational scholars.
2023 Fischer and Sitkin 361

Perhaps the classical exemplar of such a paradigm be theoretically distinct constructs that have a shared
is Mintzberg’s (1980) synthesis of research on organi- set of attributes and thus can be theoretically argued to
zational structure into five configurations: simple be recognized as coherent building blocks of a leader-
structures, machine bureaucracies, professional ship style. Specifically, there is a need for articulating
bureaucracies, divisionalized forms, and adhocra- how and why these building blocks are interrelated.
cies. These configurations, in turn, are distinguished Only if both the building blocks and their interrelation-
across two key constituent dimensions: coordinating ships are well-specified can we speak of leadership
mechanisms and design parameters. In this way, styles as meaningfully cooccurring patterns, instead of
Mintzberg’s (1980) approach is both deconflated and arbitrary and valenced laundry lists of conceptually
integrated—deconflated because he specifies config- distinct components.
urations by their constituent components (i.e., coor-
dinating mechanisms and design parameters), and Empirically Studying Leadership Styles in a
integrated because he specifies how these compo- Deconflated and Configurational Manner
nents relate to each other when forming one of the
five configurations. A reconceptualization of leadership as deconflated
More recently, Cardinal, Sitkin, and Long (2010) configurations implies reconsidering empirical meth-
synthesized research on organizational control in odologies. We now assess the prospects and limits
a similar manner. They drew on a well-established of the dominant methodology, leadership question-
set of four configurations of control types—bureau- naires, for configurational leadership research, and
cratic, clan, market, and integrative systems—and we sketch alternatives.
specified two key constituent dimensions of these: Questionnaire measures: Prospects and limits.
control mechanisms and control targets. Control Our review has shown that contemporary conceptu-
mechanisms are either formal or informal, and con- alizations of leadership styles are operationalized by
trol targets are behaviors, inputs, or outputs. Akin to questionnaire measures that have similar problems
Mintzberg’s (1980) configurations, Cardinal et al.’s in their underlying conceptualization. There is close
(2010) control types are deconflated because they are theory–method correspondence; however, it is dubi-
specified by their constituent components (i.e., con- ous if such correspondence can be maintained if
trol mechanisms and control targets), and they are leadership styles are theorized in a deconflated and
integrated because they delineate how the single configurational manner because for empirical con-
components have to relate to each other when form- figurational studies, it is important to measure not
ing a configuration. Below, we outline how leader- only the whole (i.e., leadership styles) but also its
ship research can follow the examples of Mintzberg parts (i.e., different aspects of leadership).
(1980) and Cardinal et al. (2010). We have observed that questionnaires are suscep-
Extending the configurational tradition to lead- tible to conflated measurement. Several leadership
ership styles. We suggest that a configurational questionnaires mix representations of leadership
approach can overcome the current impasse in lead- behaviors with representations of follower evalua-
ership style research. Currently, leadership styles are tions (Fischer & Dietz, 2020). Leadership question-
treated as unitary constructs or as a bundle of unitary naire items typically tap into respondents’ semantic
dimensions, despite the fact that they span multiple memory, which encodes general impressions rather
single behaviors and aspects of leadership. Thus, than concrete episodes of behaviors (Hansbrough
these types of leadership styles also need to be et al., 2021). These general impressions are likely to
decomposed in terms of their concrete, single inten- rest on multiple domains of leadership, which leads
tions, behaviors, and effects. Subsequently, there is to amalgam-based conflation. Furthermore, Eden
a need to specify whether and how these distinct and Leviatan (1975), Lord et al. (1978), and Gioia and
components form a joint construct such that it ex- Sims (1985) showed that knowledge of leadership out-
emplifies the definition of configuration as “multi- comes influences ratings of leadership styles. These
dimensional constellation of conceptually distinct results support our finding that leadership measures
characteristics that commonly occur together” (Meyer, mix distinct behavioral aspects such as behavioral
Tsui, & Hinings, 1993: 1175). content, quality of execution, or behavioral effects,
A configurational approach to studying leader- which leads to valence-based conflation.
ship styles does not imply replacing one type of Conflation in leadership questionnaires reinforces
valenced amalgam with another type of valenced the degree of subjectivity and rater specificity in
amalgam. Instead, constituent components need to measurement, such that the same leadership
362 Academy of Management Annals January

behavior might be systematically evaluated differ- represented by its constituent parts. In addition, this
ently based on the rater’s organizational role relative insight on overlaps between different leadership
to the focal leader (e.g., subordinate, peer, hierarchi- styles on evaluative aspects of leadership comple-
cal superior, or detached observer). By contrast, ments a nonconflation-based explanation for con-
should leadership questionnaires utilize noncon- struct redundancy that rests on behavioral overlap to
flated measures, their items would objectively explain high intercorrelations between measures.
describe a leader behavior, and rater specificity in Questionnaires are still useful for studying leader-
questionnaire responses would be limited to the ship styles configurationally, but have specific limi-
capacity of raters to recall a certain behavior or not. tations that must be taken into account. On one
Consequently, in the absence of systematically dif- hand, questionnaires are useful for assessing the
ferent mental recall capacities, ratings of subordi- whole—that is, leadership styles as a configuration.
nates, hierarchical superiors, peers, or detached On the other hand, using questionnaires for assess-
observers would be randomly distributed. ing the descriptive parts—the essence of displayed
However, when leadership styles conflate the leadership styles, representing patterns of leader
description with the evaluation of leadership behav- behaviors—is trickier. That is, to measure behavioral
iors, there are two sources of systematic and role- content with questionnaires in a nonconflated man-
based rater specificity in questionnaire responses. ner, items need to be formulated in a concrete and
First, the types of leadership outcomes that influence descriptive manner to avoid evaluations of behav-
questionnaire responses (see Eden & Leviatan, 1975; iors. Gioia and Sims (1985) have demonstrated that
Lord et al., 1978) likely depend on the rater’s role measures of consideration and initiating structure
relationship with the evaluated leader. For instance, lack this concreteness but that more descriptive
subordinates might rate a leader differently based on items can be developed, and that measures based on
knowledge about job satisfaction, peers based on such items are better reflections of what leaders actu-
knowledge about the performance of the leader’s ally do. However, current questionnaire measures
team, and supervisors based on knowledge about struggle to achieve this (see Table 3). Temporal separa-
compliance with organizational rules. Second, the tion of measuring behaviors and outcomes such that
semantic memory of evaluators might influence rat- responses are not influenced by performance informa-
ings in evaluator-specific ways too. For instance, sub- tion is one way of reducing, albeit not eliminating,
ordinates might have interpreted previous assertive concerns. Currently, the quality of questionnaire-
leadership behaviors as dominating, whereas supervi- based measurement is mostly judged according to psy-
sors might have intended and remembered the same chometric criteria such as discriminant or incremental
behavior as determined. Consequently, the semantic validity (e.g., Banks et al., 2016; Hoch et al., 2018);
memory of supervisors might influence future ratings though these are helpful, future research needs to also
of leadership behaviors more positively than that of assess whether questionnaires allow for nonconflated
subordinates. Jointly, these mechanisms indicate how behavioral measurement.
conceptual conflation can lead to systematic bias in Questionnaires can be used effectively for assessing
measuring leadership behaviors. the evaluative aspects of leadership. For instance,
A conflation-based explanation for why leader- attributional scholars have measured presumed leader
ship questionnaires exhibit construct redundancy intentions irrespective of actual leadership behaviors
between different leadership styles is that different (Martinko et al., 2007). For this purpose, question-
leadership questionnaires are supposed to measure naires can be used to assess attributions of intended
distinct leadership behaviors yet largely tap into the leadership styles. In addition, scholars in industrial-
same positivity or negativity in evaluating leader- organizational psychology use questionnaires to dis-
ship (i.e., valence-based conflation and overlap in tinctly evaluate the quality with which leadership
evaluative aspects of leadership). For instance, ethi- behaviors are executed and the effects they produce
cal and transformational leadership might rest on (Campbell, 2012; Campbell & Wiernik, 2015). With a
considerably different behaviors, but the measures focus on behavioral execution, questionnaires can be
of these leadership styles might capture shared posi- used for assessing realized effects of leadership styles.
tive leader evaluations rather than concrete and dif- Taken together, the current usage of questionnaire
ferent leader behaviors. Such conflated measurement measures requires either refinement (e.g., for behav-
is inconsistent with studying leadership styles config- ioral content) or explicit delimitation (e.g., for evalua-
urationally because it allows for assessing the valence tions only). Future measurement needs to proceed
of the whole but does not capture the pattern with increased accuracy when using questionnaires,
2023 Fischer and Sitkin 363

and, more generally, we need to extend our criteria on questionnaires toward a more multifaceted ap-
for judging the quality of questionnaires to avoid proach. More rigorous questionnaires, but also (com-
conflated measurement and ensure causally well- puterized) coding and qualitative case study analysis,
determined research. are exemplary methods for future leadership style
Methodological alternatives to questionnaire research.
measures. Irrespective of using more appropriate
questionnaire measures, empirical research on lead- The Future of Leadership Style Research
ership configurations would profit from more meth-
odological diversity that takes advantage of the We wish to conclude by addressing what the
strengths of other research methods. Two avenues future of leadership style research should look like.
seem to be particularly promising: (a) methods focus- The most important insights emanating from our
ing on deconflated measurement, and (b) methods meta-level review are that we need to (a) use decon-
focusing on configuring distinct components. flated leadership constructs to overcome causal
First, behavioral coding is helpful for separately indeterminacy and (b) configure these constructs to
measuring the behavioral content of leadership styles. ensure conceptual integration.
Bales (1950) pioneered such an approach for the Micro-oriented scholars are both conceptually and
study of leadership decades ago. Recently, behavioral methodologically well-equipped for advancing decon-
coding has also been used to assess the verbal and flated leadership style research because they use more
nonverbal content of charismatic leadership (Jacquart precise and nonconflated constructs and measures
& Antonakis, 2015). Accurate and descriptive coding than have been applied in extant mainstream leader-
schemes can reduce the risk of mixing the measure- ship style research (cf. Hunter et al., 2007). For
ment of behavioral content with evaluative aspects instance, rhetorical approaches to charismatic leader-
such as underlying intentions or quality of execution. ship (Antonakis et al., 2016; Baur et al., 2016; Shamir
Still, human coding is always prone to subjectivity et al., 1994) have overcome limitations of previous
and the description of behaviors can be mixed with research practice (Van Knippenberg & Sitkin, 2013;
their evaluation. Hence, computerized behavioral Yukl, 1999) while maintaining the centrality of values
coding schemes are particularly promising for non- for emerging as a charismatic leader. Another rhetori-
conflated behavioral measurement. For charismatic cal approach to studying leadership would be to exam-
leadership, Tur et al. (2021) used such an objective ine how people talk about their leadership styles,
measure. More behavioral coding, and notably com- which might be distinct from their intended leader-
puterized coding, can help future research to sepa- ship styles, displayed leadership styles, or the realized
rately measure the descriptive parts of leadership effects thereof. Drawing from other lines of micro-
configurations. oriented scholarship outside of leadership research
Second, qualitative comparative case analysis can also help move the field forward. An example is
(QCA) is helpful for empirically studying the emer- research showing how cues of competence and status
gence of leadership configurations from distinct sin- increase a person’s social influence (Berger, Ridgeway,
gle components. QCA is an established methodology Fisek, & Norman, 1998; Correll & Ridgeway, 2003);
in sociology (Schneider & Wagemann, 2010), as well such insights can reinvigorate task-oriented or instru-
as in organization theory and strategy research mental leadership research.
(Greckhamer, Misangyi, Elms, & Lacey, 2008), and Macro-oriented scholars are also well-equipped
the methodology has been put forward to study the for building on deconflated micro-oriented leader-
emergence of configurations (Fiss et al., 2013; Misan- ship research to advance a configurational approach.
gyi et al., 2017). Whereas the QCA method is mainly Scholars can examine whether different configura-
conducted at the organizational level of analysis, tions of organizational structure (Mintzberg, 1980) or
adaptations for the individual or dyadic level are organizational control (Cardinal et al., 2010) are
promising for studying leadership styles in a decon- associated with different leadership configurations
flated and reconfigured manner. (i.e., styles). In this way, scholars can also reintegrate
In summary, there are both quantitative and quali- macro-organization theory and leadership research,
tative methods with distinctive advantages for mov- which have become increasingly disconnected (Heath
ing forward a deconflated, causally well-determined, & Sitkin, 2001).
and configurational study of leadership. Future Taken together, we have the tools to successfully
research might move beyond the contemporary fix the problems identified and to avoid repeating the
methodological monoculture of relying predominantly pattern of merely replacing one conflated leadership
364 Academy of Management Annals January

style with a newer but similarly conflated style. As a Alvesson, M., & Einola, K. 2019. Warning for excessive
field, we should abstain from moving from one type positivity: Authentic leadership and other traps
of big, valenced concept (e.g., authentic, ethical, or in leadership studies. Leadership Quarterly, 30:
transformational leadership) to another type of big, 383–395.
valenced concept (e.g., caring, empathic, or inclusive Amundsen, S., & Martinsen, Ø. L. 2014. Empowering lead-
leadership), or from “excessive positivity” (Alvesson & ership: Construct clarification, conceptualization, and
Einola, 2019) to excessive negativity (e.g., harmful or validation of a new scale. Leadership Quarterly, 25:
toxic leadership) in leadership style research. Leader- 487–511.
ship style research is at a critical turning point, yet we Anderson, M. H., & Sun, P. Y. 2017. Reviewing leadership
are well-equipped for studying leadership styles in a styles: Overlaps and the need for a new “full-range” the-
deconflated and configurational way. Such a funda- ory. International Journal of Management Reviews,
mental reorientation is in order to avoid the loss of 19: 76–96.
credibility given the increasing concerns about the Antonakis, J. 2017. On doing better science: From thrill of
validity of leadership style research findings. This crit- discovery to policy implications. Leadership Quar-
ical turning point offers a unique chance for recogniz- terly, 28: 5–21.
ing these fundamental problems in a way that fosters Antonakis, J., d’Adda, G., Weber, R., & Zehnder, C. 2021.
transforming the field to become more rigorous by Just words? Just speeches? On the economic value of
embracing a de-conflated and configurational approach. charismatic leadership. Management Science. doi:
10.1287/mnsc.2021.4219
CONCLUSION Antonakis, J., Avolio, B. J., & Sivasubramaniam, N. 2003.
Context and leadership: An examination of the nine-
Our review maps the landscape of leadership style factor full-range leadership theory using the Multifac-
research and uncovers valence-based conflation as a tor Leadership Questionnaire. Leadership Quarterly,
critical limitation, leading to amalgamation, construct 14: 261–295.
redundancy, and causal indeterminacy, which makes Antonakis, J., Bastardoz, N., Jacquart, P., & Shamir, B.
it impossible to causally interpret the existing knowl- 2016. Charisma: An ill-defined and ill-measured gift.
edge base. The breadth of our review demonstrates Annual Review of Organizational Psychology and
that valence-based conflation is not the isolated weak- Organizational Behavior, 3: 293–319.
ness of one or a few leadership styles but pervades the Antonakis, J., Bendahan, S., Jacquart, P., & Lalive, R. 2010.
whole field of leadership style research due to a ten- On making causal claims: A review and recommenda-
dency to study what good and bad leaders do instead tions. Leadership Quarterly, 21: 1086–1120.
of studying what leaders do. The problem is not the Antonakis, J., & Day, D. V. 2017. The nature of leadership.
examination of good and bad leadership per se but (3rd ed.): Thousand Oaks, CA: SAGE Publications.
valence built into the research concepts. Despite this
Antonakis, J., & House, R. J. 2013. The full-range leader-
sobering assessment of the state-of-the-science, there ship theory: The way forward. In B. J. Avolio, & F. J.
are clear prospects for better leadership style research Yammario (Eds.), Transformational and charismatic
in the future. Our review identifies a nascent decon- leadership: The road ahead 10th anniversary edi-
flated and configurational approach that allows for tion, Bingley, U.K.: Emerald Group Publishing.
overcoming the identified weaknesses, and we outline Antonakis, J., & House, R. J. 2014. Instrumental leadership:
how the use of our four aspects of leadership frame- Measurement and extension of transformational–
work can guide future research toward more produc- transactional leadership theory. Leadership Quar-
tive scholarship. terly, 25: 746–771.
Arnold, J. A., Arad, S., Rhoades, J. A., & Drasgow, F. 2000.
REFERENCES The empowering leadership questionnaire: The con-
Ahearne, M., Mathieu, J., & Rapp, A. 2005. To empower or struction and validation of a new scale for measuring
not to empower your sales force? An empirical exami- leader behaviors. Journal of Organizational Behav-
nation of the influence of leadership empowerment ior, 21: 249–269.
behavior on customer satisfaction and performance. Ashford, S. J., & Sitkin, S. B. 2019. From problems to pro-
Journal of Applied Psychology, 90: 945–955. gress: A dialogue on prevailing issues in leadership
Alvesson, M. 2020. Upbeat leadership: A recipe for—or research. Leadership Quarterly, 30: 454–460.
against—“successful” leadership theory. Leadership Avolio, B. J., Bass, B. M., & Jung, D. I. 1999. Re-examining
Quarterly, 31: 945–955. doi: 10.1016/j.leaqua.2020. the components of transformational and transac-
101439 tional leadership using the Multifactor Leadership
2023 Fischer and Sitkin 365

Questionnaire. Journal of Occupational and Orga- Bedeian, A. G., & Hunt, J. G. 2006. Academic amnesia and
nizational Psychology, 72: 441–462. vestigial assumptions of our forefathers. Leadership
Avolio, B. J., & Gardner, W. L. 2005. Authentic leadership Quarterly, 17: 190–205.
development: Getting to the root of positive forms of Bedi, A., Alpaslan, C. M., & Green, S. 2016. A meta-
leadership. Leadership Quarterly, 16: 315–338. analytic review of ethical leadership outcomes
and moderators. Journal of Business Ethics, 139:
Avolio, B. J., Gardner, W. L., Walumbwa, F. O., Luthans,
517–536.
F., & May, D. R. 2004. Unlocking the mask: A look at
the process by which authentic leaders impact fol- Berger, J., Ridgeway, C. L., Fisek, M. H., & Norman, R. Z.
lower attitudes and behaviors. Leadership Quarterly, 1998. The legitimation and delegitimation of power
15: 801–823. and prestige orders. American Sociological Review,
63: 379–405.
Avolio, B. J., Walumbwa, F. O., & Weber, T. J. 2009.
Leadership: Current theories, research, and future Blake, R., & Mouton, J. 1964. The managerial grid. Hous-
directions. Annual Review of Psychology, 60: ton, TX: Gulf.
421–449. Blanchard, K. H., Zigarmi, D., & Nelson, R. B. 1993. Situa-
Bales, R. F. 1950. A set of categories for the analysis tional LeadershipV R after 25 years: A retrospective.

of small group interaction. American Sociological Journal of Leadership Studies, 1: 21–36.


Review, 15: 257–263. Bligh, M. C., Kohles, J. C., & Meindl, J. R. 2004. Charisma
Banks, G., Engemann, K., Williams, C., Gooty, J., McCau- under crisis: Presidential leadership, rhetoric, and
ley, K., & Medaugh, M. 2017. A meta-analytic review media responses before and after the September 11th
and future research agenda of charismatic leadership. terrorist attacks. Leadership Quarterly, 15: 211–239.
Leadership Quarterly, 28: 508–529. Bolino, M. C., & Grant, A. M. 2016. The bright side of being
Banks, G., Fischer, T., Gooty, J., & Stock, G. 2021a. Ethical prosocial at work, and the dark side, too: A review
leadership: Mapping the terrain for concept cleanup and agenda for research on other-oriented motives,
and a future research agenda. Leadership Quarterly, behavior, and impact in organizations. Academy of
32. doi: 10.1016/j.leaqua.2020.101471 Management Annals, 10: 599–670.
Banks, G., Gooty, J., Ross, R., Williams, C., & Harrington, Borman, W. C., & Brush, D. H. 1993. More progress toward
N. 2018. Construct redundancy in leader behaviors: A a taxonomy of managerial performance requirements.
review and agenda for the future. Leadership Quar- Human Performance, 6: 1–21.
terly, 29: 236–251.
Bowers, D. G., & Seashore, S. E. 1966. Predicting organi-
Banks, G., McCauley, K., Gardner, W., & Guler, C. 2016. A zational effectiveness with a four-factor theory of
meta-analytic review of authentic and transforma- leadership. Administrative Science Quarterly, 11:
tional leadership: A test for redundancy. Leadership 238–263.
Quarterly, 27: 634–652.
Brown, M. E., Trevi~
no, L. K., & Harrison, D. A. 2005. Ethi-
Banks, G., Woznyj, H. M., & Mansfield, C. A. 2021b. Where cal leadership: A social learning perspective for
is “behavior” in organizational behavior? A call for a construct development and testing. Organizational
revolution in leadership research and beyond. Lead- Behavior and Human Decision Processes, 97:
ership Quarterly. doi: 10.1177/17427150221132398 117–134.
Bass, B., & Avolio, B. 1995. MLQ multifactor leadership Burns, J. M. 1978. Leadership. New York, NY: Harper &
questionnaire. Menlo Park, CA: Mind Garden. Row.
Bass, B. M. 1985. Leadership and performance beyond Campbell, J. P. 2012. Behavior, performance, and effective-
expectations. New York, NY: Free Press. ness in the twenty-first century. In S. Kozlowski (Ed.),
Bass, B. M., & Bass, R. 2008. The Bass handbook of lead- The Oxford handbook of organizational psychology,
ership: Theory, research, and managerial applica- vol. 1: 159–194. Oxford, U.K.: Oxford University
tions. New York, NY: Simon and Schuster. Press.
Bass, B. M., & Steidlmeier, P. 1999. Ethics, character, and Campbell, J. P., & Wiernik, B. M. 2015. The modeling and
authentic transformational leadership behavior. Lead- assessment of work performance. Annual Review
ership Quarterly, 10: 181–217. of Organizational Psychology and Organizational
Behavior, 2: 47–74.
Baur, J. E., Ellen III, B. P., Buckley, M. R., Ferris, G. R., Alli-
son, T. H., McKenny, A. F., & Short, J. C. 2016. More Cardinal, L., Sitkin, S., & Long, C. 2010. A configurational
than one way to articulate a vision: A configurations theory of control. In S. Sitkin, L. Cardinal, & K.
approach to leader charismatic rhetoric and influence. Bijlsma-Frankema (Eds.), Organizational control.
Leadership Quarterly, 27: 156–171. New York, NY: Cambridge University Press.
366 Academy of Management Annals January

Carton, A. M. 2022. The science of leadership: A theoreti- Eisenbeiss, S. A. 2012. Re-thinking ethical leadership: An
cal model and research agenda. Annual Review interdisciplinary integrative approach. Leadership
of Organizational Psychology and Organizational Quarterly, 23: 791–808.
Behavior, 9: 61–93. Ernst, B. A. Banks, G. C. Loignon, A. C., Frear, K. A.,
Chen, G., Sharma, P. N., Edinger, S. K., Shapiro, D. L., & Williams, C. E., Arciniega, L. M., Gupta, R. Kodydek,
Farh, J.-L. 2011. Motivating and demotivating forces G. K., & Subramanian, D. 2021. Virtual charismatic
in teams: Cross-level influences of empowering lead- leadership and signaling theory: A prospective meta-
ership and relationship conflict. Journal of Applied analysis in five countries. Leadership Quarterly. doi:
Psychology, 96: 541–557. 10.1016/j.leaqua.2021.101541
Cheong, M., Yammarino, F. J., Dionne, S. D., Spain, S. M., & Eva, N., Robin, M., Sendjaya, S., van Dierendonck, D., &
Tsai, C.-Y. 2019. A review of the effectiveness of Liden, R. C. 2019. Servant leadership: A systematic
empowering leadership. Leadership Quarterly, 30: review and call for future research. Leadership Quar-
34–58. terly, 30: 111–132.
Chiniara, M., & Bentein, K. 2016. Linking servant leader- Ferris, G. R., Zinko, R., Brouer, R. L., Buckley, M. R., & Har-
ship to individual performance: Differentiating the vey, M. G. 2007. Strategic bullying as a supplemen-
mediating role of autonomy, competence and related- tary, balanced perspective on destructive leadership.
ness need satisfaction. Leadership Quarterly, 27: Leadership Quarterly, 18: 195–206.
124–141. Fischer, T., & Dietz, J. 2020. Measures of authentic, ethical,
Collins, M. D., & Jackson, C. J. 2015. A process model of and servant leadership spuriously predict objective
self-regulation and leadership: How attentional re- outcomes. Paper presented at the Academy of Man-
source capacity and negative emotions influence agement Proceedings.
constructive and destructive leadership. Leadership Fischer, T., Tian, A. W., Lee, A., & Hughes, D. J. 2021. Abu-
Quarterly, 26: 386–401. sive supervision: A systematic review and fundamen-
Correll, S. J., & Ridgeway, C. L. 2003. Expectation states tal rethink. Leadership Quarterly. doi: 10.1016/j.
theory. In Handbook of social psychology: 29–51. leaqua.2021.101540
Amsterdam, Netherlands: Kluwer Academic Publisher. Fiss, P. C. 2007. A set-theoretic approach to organizational
Delery, J. E., & Doty, D. H. 1996. Modes of theorizing configurations. Academy of Management Review,
in strategic human resource management: Tests of 32: 1180–1198.
universalistic, contingency, and configurational per- Fiss, P. C., Marx, A., & Cambr
e, B. 2013. Configurational
formance predictions. Academy of Management theory and methods in organizational research:
Journal, 39: 802–835. Introduction. Bingley, U.K.: Emerald Group
Den Hartog, D. N. 2015. Ethical leadership. Annual Publishing.
Review of Organizational Psychology and Organi- Fleishman, E. A. 1953. The description of supervisory
zational Behavior, 2: 409–434. behavior. Journal of Applied Psychology, 37: 1.
DeRue, D. S., Nahrgang, J. D., Wellman, N., & Humphrey, Fleishman, E. A., Mumford, M. D., Zaccaro, S. J., Levin,
S. E. 2011a. Trait and behavioral theories of leader- K. Y., Korotkin, A. L., & Hein, M. B. 1991. Taxonomic
ship: An integration and meta-analytic test of their rel- efforts in the description of leader behavior: A synthe-
ative validity. Personnel Psychology, 64: 7–52. sis and functional interpretation. Leadership Quar-
DeRue, D. S., Sitkin, S. B., & Podolny, J. M. 2011b. From terly, 2: 245–287.
the guest editors: Teaching leadership—Issues and Flynn, F. J., & Staw, B. M. 2004. Lend me your wallets: The
insights. Academy of Management Learning and effect of charismatic leadership on external support
Education. doi: 10.5465/amle.2011.0004 for an organization. Strategic Management Journal,
Eden, D., & Leviatan, U. 1975. Implicit leadership theory 25: 309–330.
as a determinant of the factor structure underlying Gardner, W. L., Avolio, B. J., Luthans, F., May, D. R., &
supervisory behavior scales. Journal of Applied Psy- Walumbwa, F. 2005. “Can you see the real me?” A
chology, 60: 736–741. self-based model of authentic leader and follower
Ehrhart, M. G. 2004. Leadership and procedural justice development. Leadership Quarterly, 16: 343–372.
climate as antecedents of unit-level organizational Gardner, W. L., Cogliser, C. C., Davis, K. M., & Dickens,
citizenship behavior. Personnel Psychology, 57: M. P. 2011. Authentic leadership: A review of the lit-
61–94. erature and research agenda. Leadership Quarterly,
Einarsen, S., Aasland, M. S., & Skogstad, A. 2007. Destruc- 22: 1120–1145.
tive leadership behaviour: A definition and concep- Gardner, W. L., Karam, E. P., Alvesson, M., & Einola, K.
tual model. Leadership Quarterly, 18: 207–216. 2021. Authentic leadership theory: The case for and
2023 Fischer and Sitkin 367

against. Leadership Quarterly. doi: 10.1016/j.leaqua. House, R. J. 1996. Path-goal theory of leadership: Lessons,
2021.101495 legacy, and a reformulated theory. Leadership Quar-
George, B. 2003. True north: Discover your authentic terly, 7: 323–352.
leadership. New York, NY: John Wiley & Sons. House, R. J., & Mitchell, T. R. 1975. Path-goal theory of
Gioia, D. A., & Sims, H. P., Jr. 1985. On avoiding the influ- leadership [Technical report no. 75-67]. Seattle, WA:
ence of implicit leadership theories in leader behavior University of Washington.
descriptions. Educational and Psychological Mea- Howell, J. M., & Avolio, B. J. 1992. The ethics of charis-
surement, 45: 217–232. matic leadership: Submission or liberation? Academy
Gong, Y., Huang, J.-C., & Farh, J.-L. 2009. Employee learn- of Management Executive, 6: 43–54.
ing orientation, transformational leadership, and Hunter, S. T., Bedell-Avers, K. E., & Mumford, M. D. 2007.
employee creativity: The mediating role of employee The typical leadership study: Assumptions, implica-
creative self-efficacy. Academy of Management Jour- tions, and potential remedies. Leadership Quarterly,
nal, 52: 765–778. 18: 435–446.
Greckhamer, T., Misangyi, V. F., Elms, H., & Lacey, R. Ichniowski, C., Shaw, K. L., & Prennushi, G. 1997. The
2008. Using qualitative comparative analysis in strate- effects of human resource management practices
gic management research: An examination of combina- on productivity. American Economic Review, 87:
tions of industry, corporate, and business-unit effects. 291–313.
Organizational Research Methods, 11: 695–726.
Jacquart, P., & Antonakis, J. 2015. When does charisma
Hansbrough, T. K., Lord, R. G., Schyns, B., Foti, R. J., matter for top-level leaders? Effect of attributional
Liden, R. C., & Acton, B. P. 2021. Do you remember? ambiguity. Academy of Management Journal, 58:
Rater memory systems and leadership measurement. 1051–1074.
Leadership Quarterly, 32. doi: 10.1016/j.leaqua.
Jordan, J., Brown, M. E., Trevi~
no, L. K., & Finkelstein, S.
2020.101455
2013. Someone to look up to: Executive–follower ethi-
Harris, K. J., Kacmar, K. M., & Zivnuska, S. 2007. An inves- cal reasoning and perceptions of ethical leadership.
tigation of abusive supervision as a predictor of perfor- Journal of Management, 39: 660–683.
mance and the meaning of work as a moderator of the
Judge, T. A., & Piccolo, R. F. 2004. Transformational and
relationship. Leadership Quarterly, 18: 252–263.
transactional leadership: A meta-analytic test of their
Harris, T. B., Li, N., Boswell, W. R., Zhang, X. a., & Xie, Z. relative validity. Journal of Applied Psychology, 89:
2014. Getting what’s new from newcomers: Empowering 755.
leadership, creativity, and adjustment in the socializa-
Judge, T. A., Piccolo, R. F., & Ilies, R. 2004. The forgotten
tion context. Personnel Psychology, 67: 567–604.
ones? The validity of consideration and initiating
Heath, C., & Sitkin, S. B. 2001. Big-B versus Big-O: What is structure in leadership research. Journal of Applied
organizational about organizational behavior? Journal Psychology, 89: 36.
of Organizational Behavior, 22: 43–58.
Kalshoven, K., Den Hartog, D. N., & De Hoogh, A. H. 2011.
Hill, A., Johnson, S., Greco, L., O’Boyle, E., & Walter, S. Ethical leadership at work questionnaire (ELW):
2020. Endogeneity: A review and agenda for the Development and validation of a multidimensional
methodology-practice divide affecting micro and measure. Leadership Quarterly, 22: 51–69.
macro research. Journal of Management. doi: 10.
1177/0149206320960533 Kark, R., Shamir, B., & Chen, G. 2003. The two faces
of transformational leadership: Empowerment and
Hirsch, P. M., & Levin, D. Z. 1999. Umbrella advocates ver- dependency. Journal of Applied Psychology, 88: 246.
sus validity police: A life-cycle model. Organization
Science, 10: 199–212. Katz, D., & Kahn, R. L. 1978. The social psychology of
organizations. New York, NY: Wiley.
Hoch, J. E., Bommer, W. H., Dulebohn, J. H., & Wu, D.
2018. Do ethical, authentic, and servant leadership Keller, R. T. 2006. Transformational leadership, initiating
explain variance above and beyond transformational structure, and substitutes for leadership: A longitudinal
leadership? A meta-analysis. Journal of Manage- study of research and development project team perfor-
ment, 44: 501–529. mance. Journal of Applied Psychology, 91: 202.
House, R. J. 1971. A path goal theory of leader effectiveness. Kipnis, D., Schmidt, S. M., & Wilkinson, I. 1980. Intraorga-
Administrative Science Quarterly, 16: 321–339. nizational influence tactics: Explorations in getting
one’s way. Journal of Applied Psychology, 65: 440.
House, R. J. 1977. A 1976 theory of charismatic leadership.
In J. G. Hunt & L. L. Larson (Eds.), Leadership: The Kirkman, B. L., & Rosen, B. 1999. Beyond self-management:
cutting edge. Carbondale, IL: Southern Illinois Uni- Antecedents and consequences of team empowerment.
versity Press. Academy of Management Journal, 42: 58–74.
368 Academy of Management Annals January

Kotter, J. P. 1990. Force for change: How leadership dif- Lin, S.-H., Scott, B. A., & Matta, F. K. 2019. The dark side
fers from management. New York, NY: Simon & of transformational leader behaviors for leaders them-
Schuster. selves: A conservation of resources perspective. Acad-
Krasikova, D. V., Green, S. G., & LeBreton, J. M. 2013. emy of Management Journal, 62: 1556–1582.
Destructive leadership: A theoretical review, integra- Lind, E., & Sitkin, S. 2015. The six domains leadership.
tion, and future research agenda. Journal of Manage- Dubuque, IA: Great River Learning.
ment, 39: 1308–1338.
Lord, R. G., Binning, J. F., Rush, M. C., & Thomas, J. C.
Ladkin, D., & Taylor, S. S. 2010. Enacting the “true self”: 1978. The effect of performance cues and leader
Towards a theory of embodied authentic leadership. behavior on questionnaire ratings of leadership
Leadership Quarterly, 21: 64–74. behavior. Organizational Behavior and Human Per-
Lambert, L. S., Tepper, B. J., Carr, J. C., Holt, D. T., & Bare- formance, 21: 27–39.
lka, A. J. 2012. Forgotten but not gone: An examina- Lord, R. G., Epitropaki, O., Foti, R. J., & Hansbrough,
tion of fit between leader consideration and initiating T. K. 2020. Implicit leadership theories, implicit
structure needed and received. Journal of Applied followership theories, and dynamic processing of
Psychology, 97: 913. leadership information. Annual Review of Organi-
Lee, A., Willis, S., & Tian, A. W. 2018. Empowering leader- zational Psychology and Organizational Behav-
ship: A meta-analytic examination of incremental ior, 7: 49–74.
contribution, mediation, and moderation. Journal of Lorinkova, N. M., & Perry, S. J. 2017. When is empower-
Organizational Behavior, 39: 306–325. ment effective? The role of leader-leader exchange in
Lemoine, G. J., & Blum, T. C. 2021. Servant leadership, empowering leadership, cynicism, and time theft.
leader gender, and team gender role: Testing a female Journal of Management, 43: 1631–1654.
advantage in a cascading model of performance. Per- Lovelace, J. B., Neely, B. H., Allen, J. B., & Hunter, S. T.
sonnel Psychology, 74: 3–28. 2019. Charismatic, ideological, & pragmatic (CIP)
Lemoine, G. J., Hartnell, C. A., & Leroy, H. 2019. Taking model of leadership: A critical review and agenda for
stock of moral approaches to leadership: An inte- future research. Leadership Quarterly, 30: 96–110.
grative review of ethical, authentic, and servant Luthans, F., & Avolio, B. 2003. Authentic Leadership
Leadership. Academy of Management Annals, 13: Development. In K. S. Cameron, J. E. Dutton & R. E.
148–187.
Quinn (Eds.), Positive Organizational Scholar-
Lewin, K. 1997. Resolving social conflicts and field the- ship, 241–258. San Francisco, CA: Berret-Koehler
ory in social science. Washington, DC: American Psy- Publishers.
chological Association. (Original work published
Mackey, J. D., Frieder, R. E., Brees, J. R., & Martinko, M. J.
1948.)
2017. Abusive supervision: A meta-analysis and empir-
Lewin, K., Lippitt, R., & White, R. K. 1939. Patterns of ical review. Journal of Management, 43: 1940–1965.
aggressive behavior in experimentally created “social
Martinko, M., Randolph-Seng, B., Shen, W., Brees, J.,
climates.” Journal of Social Psychology, 10: 269–299.
Mahoney, K., & Kessler, S. 2018. An examination of
Li, N., Chiaburu, D. S., Kirkman, B. L., & Xie, Z. 2013. Spot- the influence of implicit theories, attribution styles,
light on the followers: An examination of moderators and performance cues on questionnaire measures of
of relationships between transformational leadership leadership. Journal of Leadership & Organizational
and subordinates’ citizenship and taking charge. Per- Studies, 25: 116–133.
sonnel Psychology, 66: 225–260.
Martinko, M. J., Harvey, P., & Douglas, S. C. 2007. The
Liden, R. C., Wayne, S. J., Liao, C., & Meuser, J. D. 2014. role, function, and contribution of attribution theory
Servant leadership and serving culture: Influence on to leadership: A review. Leadership Quarterly, 18:
individual and unit performance. Academy of Man- 561–585.
agement Journal, 57: 1434–1452. Mayer, D. M., Aquino, K., Greenbaum, R. L., & Kuenzi, M.
Liden, R. C., Wayne, S. J., Meuser, J. D., Hu, J., Wu, J., & Liao, 2012. Who displays ethical leadership, and why does
C. 2015. Servant leadership: Validation of a short form it matter? An examination of antecedents and conse-
of the SL-28. Leadership Quarterly, 26: 254–269. quences of ethical leadership. Academy of Manage-
Liden, R. C., Wayne, S. J., Zhao, H., & Henderson, D. 2008. ment Journal, 55: 151–171.
Servant leadership: Development of a multidimen- Meindl, J. R., Ehrlich, S. B., & Dukerich, J. M. 1985. The
sional measure and multi-level assessment. Leader- romance of leadership. Administrative Science Quar-
ship Quarterly, 19: 161–177. terly, 30: 78–102.
Likert, R. 1961. New patterns of management. New York, Merriam-Webster. 2020.Leadership style. Retrieved from
NY: McGraw Hill. https://www.merriam-webster.com/dictionary/style
2023 Fischer and Sitkin 369

Meslec, N., Curseu, P. L., Fodor, O. C., & Kenda, R. 2020. Ong, W. J., & Johnson, M. D. 2021. Towards a configural
Effects of charismatic leadership and rewards on theory of job demands and resources. Academy of
individual performance. Leadership Quarterly, 31: Management Journal. doi: 10.5465/amj.2020.0493
101423. doi: 10.1016/j.leaqua.2020.101423 Padilla, A., Hogan, R., & Kaiser, R. B. 2007. The toxic trian-
Meyer, A. D., Tsui, A. S., & Hinings, C. R. 1993. Configura- gle: Destructive leaders, susceptible followers, and
tional approaches to organizational analysis. Acad- conducive environments. Leadership Quarterly, 18:
emy of Management Journal, 36: 1175–1195. 176–194.
Miller, D. 1986. Configurations of strategy and structure: Pfeffer, J. 1977. The ambiguity of leadership. Academy of
Towards a synthesis. Strategic Management Journal, Management Review, 2: 104–112.
7: 233–249.
Pfeffer, J. 2015. Leadership BS: Fixing workplaces and
Miller, D. 1996. Configurations revisited. Strategic Man- careers one truth at a time. New York, NY: Harper
agement Journal, 17: 505–512. Collins.
Miller, D. 2018. Challenging trends in configuration Podsakoff, P. M., MacKenzie, S. B., Moorman, R. H., & Fet-
research: Where are the configurations? Strategic ter, R. 1990. Transformational leader behaviors and
Organization, 16: 453–469. their effects on followers’ trust in leader, satisfaction,
Mintzberg, H. 1973. The nature of managerial work. New and organizational citizenship behaviors. Leadership
York, NY: Harper & Row. Quarterly, 1: 107–142.
Mintzberg, H. 1980. Structure in 5’s: A synthesis of the Popper, K. 1959. The logic of scientific discovery.
research on organization design. Management Sci- London, U.K.: Hutchinson.
ence, 26: 322–341.
Price, T. L. 2003. The ethics of authentic transformational
Mintzberg, H. 2009. Managing. San Francisco, CA: Ber- leadership. Leadership Quarterly, 14: 67–81.
rett-Koehler Publishers.
Rousseau, D. M., Manning, J., & Denyer, D. 2008. Evidence
Mintzberg, H., & Waters, J. A. 1985. Of strategies, deliberate in management and organizational science: Assem-
and emergent. Strategic Management Journal, 6: bling the field’s full weight of scientific knowledge
257–272. through syntheses. Academy of Management Annals,
Mirabeau, L., Maguire, S., & Hardy, C. 2018. Bridging prac- 2: 475–515.
tice and process research to study transient manifesta- Sajons, G. B. 2020. Estimating the causal effect of mea-
tions of strategy. Strategic Management Journal, 39: sured endogenous variables: A tutorial on experimen-
582–605. tally randomized instrumental variables. Leadership
Misangyi, V. F., Greckhamer, T., Furnari, S., Fiss, P. C., Quarterly, 31: 101348.
Crilly, D., & Aguilera, R. 2017. Embracing causal com- Schneider, C. Q., & Wagemann, C. 2010. Standards of good
plexity: The emergence of a neo-configurational per- practice in qualitative comparative analysis (QCA)
spective. Journal of Management, 43: 255–282. and fuzzy-sets. Comparative Sociology, 9: 397–418.
Morgeson, F. P., DeRue, D. S., & Karam, E. P. 2010. Leader- Schriesheim, C., & Kerr, S. 1974. Psychometric properties
ship in teams: A functional approach to understand-
of the Ohio State leadership scales. Psychological
ing leadership structures and processes. Journal of
Bulletin, 81: 756.
Management, 36: 5–39.
Schyns, B., & Schilling, J. 2013. How bad are the effects of
Neider, L. L., & Schriesheim, C. A. 2011. The authentic
bad leaders? A meta-analysis of destructive leadership
leadership inventory (ALI): Development and empiri-
and its outcomes. Leadership Quarterly, 24: 138–158.
cal tests. Leadership Quarterly, 22: 1146–1164.
Shamir, B., Arthur, M. B., & House, R. J. 1994. The rhetoric
Neubert, M. J., Hunter, E. M., & Tolentino, R. C. 2016. A
of charismatic leadership: A theoretical extension, a
servant leader and their stakeholders: When does
case study, and implications for research. Leadership
organizational structure enhance a leader’s influence?
Quarterly, 5: 25–42.
Leadership Quarterly, 27: 896–910.
Shamir, B., & Eilam, G. 2005. “What’s your story?” A life-
Neubert, M. J., Kacmar, K. M., Carlson, D. S., Chonko,
stories approach to authentic leadership development.
L. B., & Roberts, J. A. 2008. Regulatory focus as a medi-
Leadership Quarterly, 16: 395–417.
ator of the influence of initiating structure and servant
leadership on employee behavior. Journal of Applied Shamir, B., House, R. J., & Arthur, M. B. 1993. The moti-
Psychology, 93: 1220. vational effects of charismatic leadership: A self-
Ng, T. W. H., & Feldman, D. C. 2015. Ethical leadership: concept based theory. Organization Science, 4:
Meta-analytic evidence of criterion-related and incre- 577–594.
mental validity. Journal of Applied Psychology, 100: Shaw, J. B., Erickson, A., & Harvey, M. 2011. A method for
948–965. measuring destructive leadership and identifying
370 Academy of Management Annals January

types of destructive leaders in organizations. Leader- reputation for ethical leadership. California Manage-
ship Quarterly, 22: 575–590. ment Review, 42: 128–142.
Siangchokyoo, N., Klinger, R. L., & Campion, E. D. 2020. Tur, B., Harstad, J., & Antonakis, J. 2021. Effect of charis-
Follower transformation as the linchpin of transforma- matic signaling in social media settings: Evidence
tional leadership theory: A systematic review and from TED and Twitter. Leadership Quarterly, 33.
future research agenda. Leadership Quarterly, 31: doi: 10.1016/j.leaqua.2020.101476
101341.
Van Knippenberg, D., & Sitkin, S. B. 2013. A critical assess-
Sparrowe, R. T. 2005. Authentic leadership and the narra- ment of charismatic-transformational leadership re-
tive self. Leadership Quarterly, 16: 419–439. search: Back to the drawing board? Academy of
Stogdill, R. M. 1950. Leadership, membership and organi- Management Annals, 7: 1–60.
zation. Psychological Bulletin, 47: 1. Vroom, V. H., & Jago, A. G. 1988. The new leadership:
Stogdill, R. M. 1962. Manual for the LBDQ-Form XII. Managing participation in organizations. Hoboken,
Columbus, OH: Fisher College of Business, The Ohio NJ: Prentice-Hall.
State University. Waldkirch, M., Kammerlander, N., & Wiedeler, C. 2021.
Stouten, J., van Dijke, M., Mayer, D. M., De Cremer, D., & Configurations for corporate venture innovation:
Euwema, M. C. 2013. Can a leader be seen as too ethi- Investigating the role of the dominant coalition. Jour-
cal? The curvilinear effects of ethical leadership. nal of Business Venturing, 36. doi: 10.1016/j.jbus-
Leadership Quarterly, 24: 680–695. vent.2021.106137
Strange, J. M., & Mumford, M. D. 2005. The origins of Walter, F., & Bruch, H. 2009. An affective events model of
vision: Effects of reflection, models, and analysis. charismatic leadership behavior: A review, theoretical
Leadership Quarterly, 16: 121–148. integration, and research agenda. Journal of Manage-
Sun, P. Y. 2013. The servant identity: Influences on the ment, 35: 1428–1452.
cognition and behavior of servant leaders. Leadership Walumbwa, F. O., Avolio, B. J., Gardner, W. L., Wernsing,
Quarterly, 24: 544–557. T. S., & Peterson, S. J. 2008. Authentic leadership:
Development and validation of a theory-based mea-
Tannenbaum, R., & Schmidt, W. H. 1973. How to choose
sure. Journal of Management, 34: 89–126.
a leadership pattern. Harvard Business Review.
Retrieved from https://hbr.org/1973/05/how-to-choose- Walumbwa, F. O., Hartnell, C. A., & Oke, A. 2010. Servant
a-leadership-pattern. leadership, procedural justice climate, service cli-
mate, employee attitudes, and organizational citizen-
Tengblad, S. 2003. Classic, but not seminal: Revisiting the
ship behavior: A cross-level investigation. Journal of
pioneering study of managerial work. Scandinavian
Applied Psychology, 95: 517.
Journal of Management, 19: 85–101.
Wang, H., Law, K. S., Hackett, R. D., Wang, D., & Chen,
Tepper, B. J. 2000. Consequences of abusive supervision.
Z. X. 2005. Leader–member exchange as a mediator of
Academy of Management Journal, 43: 178–190.
the relationship between transformational leadership
Tepper, B. J. 2007. Abusive supervision in work organiza- and followers’ performance and organizational citi-
tions: Review, synthesis, and research agenda. Journal zenship behavior. Academy of Management Journal,
of Management, 33: 261–289. 48: 420–432.
Tepper, B. J., Henle, C. A., Lambert, L. S., Giacalone, R. A., Weber, M. 2002. Wirtschaft und Gesellschaft: Grundriss
& Duffy, M. K. 2008. Abusive supervision and subor- der verstehenden Soziologie (5th ed.). T€
ubingen, Ger-
dinates’ organization deviance. Journal of Applied many: Mohr Siebeck. (Original work published 1922.)
Psychology, 93: 721.
Willmott, H. 1984. Images and ideals of managerial work: A
Tepper, B. J., Simon, L., & Park, H. M. 2017. Abusive super- critical examination of conceptual and empirical
vision. Annual Review of Organizational Psychol- accounts. Journal of Management Studies, 2: 349–368.
ogy and Organizational Behavior, 4: 123–152.
Wofford, J., & Liska, L. Z. 1993. Path-goal theories of lead-
Thoroughgood, C. N., Padilla, A., Hunter, S. T., & Tate, ership: A meta-analysis. Journal of Management, 19:
B. W. 2012. The susceptible circle: A taxonomy of fol- 857–876.
lowers associated with destructive leadership. Lead- Yu, L., Duffy, M. K., & Tepper, B. J. 2018. Consequences of
ership Quarterly, 23: 897–917. downward envy: A model of self-esteem threat, abusive
Tierney, P., & Tepper, B. J. 2007. Introduction to The Lead- supervision, and supervisory leader self-improvement.
ership Quarterly special issue: Destructive leadership. Academy of Management Journal, 61: 2296–2318.
Leadership Quarter, 18: 171–173. Yukl, G. 1999. An evaluation of conceptual weaknesses in
Trevi~
no, L. K., Hartman, L. P., & Brown, M. 2000. Moral transformational and charismatic leadership theories.
person and moral manager: How executives develop a Leadership Quarterly, 10: 285–305.
2023 Fischer and Sitkin 371

Yukl, G. 2012. Effective leadership behavior: What we


know and what questions need more attention. Acad-
emy of Management Perspectives, 26: 66–85.
Yukl, G., & Falbe, C. M. 1990. Influence tactics and objec- Thomas Fischer (thomas.fischer@unige.ch) is an assistant
tives in upward, downward, and lateral influence professor of responsible leadership at the University
attempts. Journal of Applied Psychology, 75: 132. of Geneva. His research focuses mainly on the
conceptualization and measurement of leadership. He also
Yukl, G., & Gardner, W. L. 2019. Leadership in organiza- studies how people talk about their leadership and how
tions (9th ed.). London, U.K.: Pearson Education. they adapt their leadership behaviors to interpersonal and
Zhang, X., & Bartol, K. M. 2010. Linking empowering lead- structural demands.
ership and employee creativity: The influence of psy-
chological empowerment, intrinsic motivation, and Sim B. Sitkin (sim.sitkin@duke.edu) is Michael W.
creative process engagement. Academy of Manage- Krzyzewski University Distinguished Professor and
ment Journal, 53: 107–128. professor of management and public policy at Duke
University. His research focuses on the effects of leadership
Zhang, Y., & Liao, Z. 2015. Consequences of abusive super-
and organizational control on trust, risk-taking, courage,
vision: A meta-analytic review. Asia Pacific Journal
experimentation, learning, and innovation. His most recent
of Management, 32: 959–987.
books are Organizational Control, The Six Domains of Lead-
Zhang, Y., Liu, X., Xu, S., Yang, L.-Q., & Bednall, T. C. 2019. ership and The Routledge Companion to Trust.
Why abusive supervision impacts employee OCB and
CWB: A meta-analytic review of competing mediating
mechanisms. Journal of Management, 45: 2474–2497.
372 Academy of Management Annals January

APPENDIX A

TABLE A1
Sample of Articles Underlying the Review
Cut-Off Valueb Number of
(For Inclusion Articles
Total Number of Based on Additionally
Leadership Articles Retained Number of Articles Retained Absolute Included in the
Stylea Search Terms for Each Type Based on Inclusion Criteria Citations) Review

Positive leadership styles


Authentic authentic AND leader Conceptual: 10 Review: 6 91 3
leadership Empirical: 4 Relative number of citations (i.e.,
Review: 6 relative to the publication
year): 2
Absolute number of citations: 12
Consideration and consideration AND Conceptual: 0 Overall: 4 — 4
initiating initiating structure
Empirical: 3
structure Review: 1
Empowering empowering AND Conceptual: 0 Review: 1 22 2
leadership leader Empirical: 19 Relative number of citations: 1
Review: 1 Absolute number of citations: 18
Ethical leadership ethical AND leader Conceptual: 2 Review: 5 60 2
Empirical: 13 Relative number of citations: 6
Review: 5 Absolute number of citations: 9
Instrumental Instrumental AND Conceptual: 0 Overall: 2 — 6
leadership leader Empirical: 2
Review: 0
Servant leadership servant AND leader Conceptual: 1 Review: 4 1 1
Empirical: 15 Relative number of citations: 3
Review: 4 Absolute number of citations: 13
Transformational transformational AND Conceptual: 0 Review: 5 427 0
leadership leader Empirical: 15 Relative number of citations: 2
Review: 5 Absolute number of citations: 13
Charismatic charisma AND leader Conceptual: 2 Review: 4 79 4
leadership Empirical: 14 Relative number of citations: 0
Review: 4 Absolute number of citations: 16
Negative leadership styles
Abusive abusive AND Conceptual: 0 Review: 5 201 0
supervision (supervision OR Empirical: 15 Relative number of citations: 5
supervisor) Review: 5 Absolute number of citations: 10
Destructive destructive AND Conceptual: 4 Review: 4 — 0
leadership leader Empirical: 3 Relative number of citations: 0
Review: 4 Absolute number of citations: 7

a
For three leadership styles, we identified fewer than 20 articles. We still include these styles in the review because they have been
very influential in the past, have clear conceptualizations, and include very influential articles up to the current time (i.e., consideration
and initiating structure leadership, path–goal theory or instrumental leadership), or shed light on the dark side of leadership (i.e.,
destructive leadership). Together, the ensemble of reviewed leadership styles covers much of the ongoing research.
b
The cut-off value refers to the minimum number of citations in Web of Science that an article needs to have in order to be retained based
on the absolute number of citations. For instance, for authentic leadership the cut-off value is 91. This means that we retained all articles
published since 2000 in our set of 12 journals that have at least one of the following three features: (a) being a review article, (b) being marked
as highly cited in Web of Science (i.e., relative number of citations; top 1% for the publication year), or (c) having at least 91 citations. A cut-
off value of “—” means that there are maximally 20 identified articles; thus, all identified articles are included in the review.

You might also like