True to what We stand for: Championing collective interests as a path to authentic leadership

Niklas K. Steffens a,⁎, Frank Mols b, S. Alexander Haslama, Tyler G. Okimoto c
a. School of Psychology, The University of Queensland, Brisbane, QLD 4072, Australia
b. School of Political Science and International Studies, The University of Queensland, Brisbane, QLD 4072, Australia
c. UQ Business School, The University of Queensland, Brisbane, QLD 4072, Australia

a b s t r a c t

Growing evidence points to the role of authentic leadership in enhancing followership. Yet little
is known about the factors that determine whether followers perceive leaders as displaying
authentic leadership. In the present research, we examine the impact of leaders’ championing
of collective (group) interests on authentic leadership.

Study 1 shows experimentally that compared
to a leader who advances personal interests, a leader who advances the interests of a
collective is (a) perceived as offering more authentic leadership and (b) more likely to inspire
followership. Findings are followed up in a field study revealing that leaders’ championing of
collective interests is associated with greater perceived authentic leadership and followership
(in terms of voting intentions). Furthermore, results indicate that shared self-categorization is
a boundary condition of these relationships such that the relationship between a leader’s
championing of collective (group) interests and authentic leadership (and followership) is
more pronounced for perceivers who self-categorize as members of the group that a leader
is leading (rather than of a different group). In sum, findings suggest that leaders are regarded
as more authentic to the extent that they are true to the collective identity of the group that
they lead.

Introduction

The Chairman and CEO of the pharmaceutical company Novartis, Dan Vasella, has been lauded as a highly successful CEO
whose success is often described as being intertwined with his willingness to “follow his heart” and “walk the walk”, as exhibited
through his compassionate attempts to help those suffering from life-threatening diseases (George, 2007). Similarly, Myanmar’s
first democratically elected stateswoman, Aung San Suu Kyi, has been heralded as a leader who passionately stands up for her
firm beliefs in freedom and peaceful social change. Recent theory and research demonstrate compellingly that leaders like Vasella
and Suu Kyi who are seen as “authentic” are able to inspire a range of positive outcomes — for followers, for groups, for organizations,
and for leaders themselves (for a comprehensive review, see Gardner, Cogliser, Davis, & Dickens, 2011). Yet despite significantly
advancing our understanding of the consequences of authentic leadership, researchers have devoted less energy to
investigating the factors that shape followers’ perceptions of authentic leadership. What specific behaviors lead people to view

leaders as authentic? Such an understanding is critical both for practical development of authentic leadership, as well as for a theoretical
account of the effects of authenticity.
According to authentic leadership theory (Avolio, Gardner, Walumbwa, Luthans, & May, 2004; Gardner et al., 2011; Luthans &
Avolio, 2003), leaders are deemed authentic when their actions are seen as consistent with their personal values and beliefs; and
it is this authenticity that inspires positive outcomes. However, a leader’s values and beliefs also tend to stand for the vision and
mission of the group and the organization they lead; as such, an authentic leader can be true to both him- or herself and broader
organizational values. For example, Vasella’s compassion for people in need and his goal of substantially increasing investments in
pharmaceutical innovations might reflect his personal values and business philosophy as shaped by his personal experiences of
suffering during his childhood. Yet this might also reflect his engagement with the broader vision and mission of Novartis as
the company has evolved (a mission that distinguished the company from rival companies): “As CEO, I have the leverage to impact
the lives of many more people. At the end of the day the only thing that matters is what we do or omit to do for other people”
(cited in George, 2007, p. 49). Likewise, Suu Kyi’s willingness to champion the cause of freedom and democracy might arise
from her personal experiences and beliefs. However, it might also reflect the will and interests of the Burmese people (in contrast
to those of the Burmese military). Indeed, as she put it: “The best way to help Burma is to empower the people of Burma, to help
us have enough self-confidence to obtain what we want for ourselves” (Globalpost, 2010). In other words, perceived leader authenticity
may follow from leaders being authentic to their own beliefs, and/or their being authentic to a broader group vision
that they represent (and that differentiate their ingroup from rival outgroups). Moreover, both may be embodied in followers’
perceptions of a leader’s relationship to the collective identity, and of their actions in relation to that identity.
In the present research, we use self-categorization theory as a theoretical basis for understanding authentic leadership perceptions
as potentially originating from a leader’s pursuit of both individual and collective interests. A core tenet of this theory is that
a person’s sense of self is flexible and, depending on features of the prevailing social context, can be defined at different levels of
abstraction (Ashforth, Harrison, & Corley, 2008; after Ashforth & Mael, 1989; Turner, 1982; Turner, Hogg, Oakes, Reicher, &
Wetherell, 1987). At a lower level of abstraction, a person’s sense of personal identity (as “I”) is defined exclusively in terms of
idiosyncratic traits and attributes; at a higher level of abstraction, a sense of social identity (as “us”; e.g., “us leadership scholars”)
is defined inclusively in terms of characteristics (e.g., norms, values, and goals) that are believed to be shared with fellow ingroup
members (for reviews, see Ellemers, 2012; Postmes & Branscombe, 2010; Reicher, Spears, & Haslam, 2010).
Integrating ideas from work on authentic leadership and self-categorization theory, we propose that followers’ perceptions of
leaders’ authenticity are partly shaped by followers’ perceptions of the leader’s relationship to a collective identity, and followers’
own relationship to that identity. More specifically, we propose that followers’ perceptions of leaders’ authenticity will vary as a
function of the degree to which leaders are seen to enact a collective self by advancing the collective interests of their ingroup
(i.e., so that they are seen to be championing that group rather than other groups or themselves as individuals; Giessner, van
Knippenberg, van Ginkel, & Sleebos, 2013; Haslam, Reicher, & Platow, 2011; van Knippenberg & van Knippenberg, 2005). Thus,
it is not only personal self-consistency per se that inspires authenticity judgements, but also collective self-consistency. We first
test the proposed model in an experimental study to address issues of causality and then follow up the results in a second
study to examine these relationships in the field.
The research aims to make at least three important contributions to the literatures on leadership and followership. First, our
goal is to shed light on the origins of authentic leadership perceptions. As noted above, this is important because we know relatively
little about the factors that shape these perceptions. Second, we attempt to expand upon suggestions that authentic leadership
is affected by followers’ perceptions of leaders’ relationship to their inner (personal) self by providing evidence of the
extent to which authentic leadership is also affected by followers’ perceptions of leaders’ orientation to a collective (group)
self. Third, by examining authentic leadership perceptions and followership simultaneously, we seek to identify the conditions
under which these two processes are (and are not) aligned.
Conceptualizations and antecedents of authentic leadership
In the last decade, the impact that authentic leadership has on followers and on organizations has been a major research focus
for the leadership literature. Although authentic leadership has been defined in a range of different ways, conceptualizations center
on issues of truth and fidelity to the self, as captured in phrases such as “saying what one means” and “being true to yourself”.
The model by Avolio and colleagues integrates early approaches to authentic leadership, offering what is now the most widely
tested conceptualization of authentic leadership (Avolio & Gardner, 2005; Walumbwa, Avolio, Gardner, Wernsing, & Peterson,
2008). They conclude that “the essence of authenticity is to know, accept, and remain true to one’s self” (Avolio et al., 2004,
p. 802).
The majority of research has focused on discussing authentic leadership as a general factor (Gardner et al., 2011). Nonetheless,
various models propose that this general factor consists of lower-level sub-components. For example, the widely used model of
Walumbwa et al. (2008) (see also Neider & Schriesheim, 2011) considers four sub-components that together make up the general
factor of authentic leadership. These are: (1) self-awareness (i.e., being aware of the way in which one derives meaning from the
world and of the impact one has on others), (2) relational transparency (i.e., presenting one’s authentic, rather than fake or
distorted, self to other people as well as minimizing displays of inappropriate emotions), (3) balanced processing
(i.e., objectively analyzing all data before making decisions as well as soliciting feedback that challenges one’s own views), and
(4) internalized moral perspective (i.e., being guided by moral standards and behaving in ways that are congruent with internalized
values rather than yielding to group, organizational, and societal pressures).

Empirical evidence from a range of samples, sectors, and countries provides support for the idea that leaders who are seen
to be authentic (in ways consistent with the above definition) are also effective in the sense that they are able to mobilize followers’
energies and to increase group and organizational success. For instance, empirical evidence indicates that followers’ perceptions
of authentic leadership are positively related not only to (a) organizational performance (Clapp-Smith, Vogelgesang, &
Avey, 2009), but also to followers’ (b) job satisfaction and work engagement (Giallonardo, Wong, & Iwasiw, 2010;Wong, Spence
Laschinger, & Cummings, 2010), (c) performance (Wong & Cummings, 2009), and (d) satisfactionwith their leader (Walumbwa
et al., 2008).
As Gardner et al. (2011) observe in their comprehensive review of research on this topic, our awareness of the importance of
authentic leadership has grown dramatically since the construct was first formally explored more than a decade ago (Avolio &
Gardner, 2005; Avolio et al., 2004; Bass & Steidlmeier, 1999; Gardner, Avolio, Luthans, May, & Walumbwa, 2005; George, 2007;
Ilies, Morgeson, & Nahrgang, 2005; Luthans & Avolio, 2003; Walumbwa et al., 2008). However, despite this progress, only two
of the 91 articles identified in Gardner et al.’s (2011) review involved empirical investigation of the dynamics that are associated
with authentic leadership perceptions. In the first of these two studies Jensen and Luthans (2006) found that leaders who had
more psychological capital (i.e., greater hope, optimism, and resilience) perceived themselves to have greater authentic leadership.
In the second, Tate (2008) found that leaders’ engagement in self-monitoring was unrelated to their perceptions of their own authentic
leadership. Critically, though, none of the available research investigated antecedents of followers’ perceptions of authentic
leadership. This is an important gap in understanding since ultimately it is followers’ understanding (including their perceptions
of leaders as more or less authentic) and their subsequent actions that account for leaders’ impact in the world at large (Bennis,
1999; Haslam & Platow, 2001; Hollander, 1992).
Despite earlier calls to attend to this issue (Fields, 2007; see also Cooper, Scandura, & Schriesheim, 2005; Shamir & Eilam,
2005), this picture has not changed substantially in recent years. Nevertheless, research is starting to address this issue (and to
provide insights into underlying process) by showing that perceptions of authentic leadership are positively related to
(a) followers’ sense that there is alignment between their own personal values and the core personal values of the leader as
well as those followers’ general belief in social change (Williams, Pillai, Deptula, & Lowe, 2012), (b) leaders’ self-consistency
and knowledge of their personal self (Peus, Wesche, Streicher, Braun, & Frey, 2012), and (c) correspondence between leaders’
physical actions and both their verbal expressions and their portrayed life story (Weischer, Weibler, & Petersen, 2013). Nevertheless,
to date, little attempt has been made to integrate these various insights with literature that focuses on the role of groups in
leadership and followership processes. As we will outline in more detail below, in the present research we aim to shed light on
those group dynamics that are likely to be more (or less) conducive to authentic leadership perceptions. To this end, we examine
the contribution of principles formulated within both charismatic leadership research and the social identity approach to leadership
that focus on leaders’ championing of group interests.
Leaders’ championing of collective interests
Charismatic leadership research and social identity theorizing both make the point that leaders’ ability to influence followers
results from their being perceived to promote the interests of a higher-order collective rather than merely their personal interests.
In this regard, charismatic leadership research suggests that leaders can enhance their charisma not only through personalized
behavior (that binds followers to them as individuals) but also through socialized actions (that bind followers to a collective of
which they are representative; Shamir, House, & Arthur, 1993). The key idea here is that by transcending their personal interests
through attention to the greater good, “socialized charismatic” leaders are able to stimulate positive follower behavior and also
advance societal goals (Conger & Kanungo, 1998; Howell & Shamir, 2005). Indeed, evidence presented by Kark, Shamir, and
Chen (2003) shows that followers’ perceptions of leaders’ transformational behaviors are associated not only with followers’ personal
identification with leaders but also with followers’ identification with a collective. Moreover, evidence indicates that the former
is positively associated with followers’ dependency while the latter is positively associated with followers’ sense of
empowerment. This suggests that for leaders to be effective, they need to act (and be seen to act) in a selfless or self-sacrificial
manner that promotes interests associated with a higher-order or “greater” good.
Along similar lines, the social identity approach to leadership asserts that when people categorize themselves and others in
terms of a shared social identity (as “us”), this has important consequences for social, organizational, and political behavior in
general (Ashforth & Mael, 1989; Haslam, 2001/2004; Hogg & Terry, 2001; Turner et al., 1987; Tyler & Blader, 2003; van Dick,
2001) but for leadership and followership in particular (Ellemers, de Gilder, & Haslam, 2004; Hogg, 2001; Platow, Haslam,
Reicher, & Steffens, 2014; Thomas, Martin, & Riggio, 2013; Turner, 1991; Turner & Haslam, 2001; van Knippenberg & Hogg,
2003; van Knippenberg, 2011). Among other things, the approach asserts that in order to succeed in influencing followers, leaders
need to act as identity champions by advancing the collective interests of the (in)group that they are leading rather than their
personal interests or those of other (out)groups (Giessner et al., 2013; Graf, Schuh, van Quaquebeke, & van Dick, 2012; Haslam
& Platow, 2001; Haslam et al., 2001, 2011; Steffens, Haslam, Reicher, Platow et al., 2014b; van Dick, Hirst, Grojean, & Wieseke,
2007; van Knippenberg & van Knippenberg, 2005).
This idea of leaders as champions of a group also has some similarities with the notion of leaders as group prototypes
(i.e., people who embody the norms, beliefs, and values that characterize a group; for reviews, see Haslam et al., 2011; Hogg,
van Knippenberg, & Rast, 2012; van Knippenberg, 2011). While leaders’ prototypicality and championing of collective interests
will often go hand in hand, they are not the same and can also diverge. Specifically, while leaders who are seen as prototypical
(or representative) of an identity may generally also engage in behavior that champions interests associated with that identity,

it is possible (a) for highly prototypical leaders to be unwilling to engage in behavior that involves standing up for the group and
actively furthering collective interests, and (b) for leaders who are not prototypical of a group to behave in ways that actively
champion the interests of that group (for more in-depth discussion of similarities and differences between these two concepts,
see Haslam et al., 2011; Steffens et al., 2014b; Steffens, Schuh, Haslam, Pérez, & Dick, 2015). Moreover, it is also noteworthy
that in contrast to transformational models, the social identity approach argues that for leaders to be effective, they need to
forgo their personal interests not for any higher-order group but rather for interests that are specific to a salient (in)group that
is self-defining for followers.
At first glance, the suggestion that leadership involves sacrificing personal interests for group interests, emphasized in both
charismatic and social identity approaches, appears to be at odds with the idea that effective leaders are authentically “true to
one’s self”. However, we argue that these viewpoints are actually complementary. Building on the simple but important idea
that leaders may construct their sense of self on the basis of who they are as distinct individuals as well as on the basis of
who they are as members of a distinct collective, we propose that followers may make sense of leaders’ authenticity not only
with reference to leaders’ personal self but also with reference to a collective self — that is, the group or higher-order entity
that they are leading. Moreover, one important way in which followers may understand whether leaders are being true to
such a collective self is through their assessment that the leader is standing up for that collective and championing its collective
interests.
By way of example, one would expect that an action by a national leader will be seen as more authentic when she or he stands
up for the nation through actions that further its cause rather than the cause of other nations or themselves as individuals. Thus
U.S. President George W. Bush’s authentic leadership attracted fierce criticism in 2012 for charging Helping a Hero (a US war
veteran’s charity) $100,000 for a speech he delivered, but he was seen as more authentic (more self-aware, more relationally
transparent, more balanced, more moral) a decade earlier, when he stood up for America’s interests in the wake of the 9/11 attacks
(New York Times, 2001). Indeed, this example resonates with evidence provided by Bligh, Kohles, and Meindl (2004) that
Bush’s rhetoric in the aftermath of 9/11 made (and was portrayed by the media as making) stronger appeals to general values.
Similarly, there were few objections to the authentic leadership of Australian Prime Minister Tony Abbott when he commended
Australian troops for their bravery and for embodying the ANZAC spirit, but he was seen as inauthentic when he awarded an
Australian Knighthood — an honor that is believed to serve the purpose of celebrating Australians of extraordinary achievement
and merit — to Britain’s Prince Philip (Guardian, 2015). More formally, this argument leads to the following hypothesis:
H1. Followers will regard a leader as displaying authentic leadership to the extent that that leader is seen to champion the collective
interests of the group that he or she is leading rather than personal interests or those of an outgroup.
Yet beyond perceptions of authentic leadership, it is also instructive to reflect on the simultaneous impact that the
championing of collective interests has on followership, with a view to gaining a better understanding of the extent to which perceptions
of authentic leadership and followership go hand in hand (or else diverge). In this regard, theory and research suggest
that advancing collective interests will have a bearing not just on perceptions of leadership but also on acts of followership. For
example, an experiment by Haslam and Platow (2001) showed that when leaders stood up for their own group (rather than a
competing outgroup) followers generated more creative ideas to help turn the leaders’ plans into reality (see also Haslam
et al., 2001). Similarly, van Knippenberg and van Knippenberg (2005) showed that leaders who sacrifice their own interests for
those of their group are seen to be more effective and proven more capable of enhancing follower performance. Moreover, several
experimental and field studies by van Dick and colleagues show that leaders who are seen to identify strongly with their group
(and by inference who are more willing to advance the interests of that group because they see the group as self-defining) are
more effective, as indicated by their capacity to increase followers’ (a) identification with an organization and organizational citizenship
(van Dick et al., 2007), (b) performance (van Dick & Schuh, 2010), and (c) personal identification with the leader
(Steffens et al., 2015).
Together, these various strands of research suggest that leaders’ group-oriented behavior is associated with followers’ positive
response to those leaders. In the present research, we seek to expand upon this insight by examining the degree to which follower
perceptions of political leaders’ championing of collective interests affect their followership as defined by voting intentions.
Here we anticipate that followers will be more willing to follow a leader by voting for him or her when that leader is seen to
be true to the collective self in championing interests associated with the group that they are leading. More specifically, we hypothesize
that:
H2. Followers will be more likely to follow a leader (as assessed by intentions to vote for the leader) to the extent that the leader
is seen to champion collective interests of the group that he or she is leading, rather than personal interests or those of an
outgroup.
The preceding discussion suggests that leaders’ championing of collective interests may be a path both to authentic leadership
perceptions and to followership. Does this mean, though, that leaders only need to advance the interests of their collective constituency
to be regarded as inherently true and authentic? Or does leaders’ championing of collective interests promote a sense of
authenticity in the eyes of some followers but not others? In this regard, the literature suggests that followers’ responses are likely
to vary as a function of whether or not they share collective identity with the leader (Turner, 1991; for empirical evidence, see
McGarty, Haslam, Hutchinson, & Turner, 1994; Platow & van Knippenberg, 2001; Steffens, Haslam, & Reicher, 2014a; Ullrich,
Christ, & van Dick, 2009; Van Dijke & De Cremer, 2010). More specifically, we suggest that a leader’s enactment and advancement

of collective interests should have a stronger impact on followers’ perceptions of leaders’ authenticity when followers perceive
leaders to be aligned with a collective entity that they see as self-defining.
It is also worth highlighting some of the key conceptual similarities and differences between, on the one hand, shared group
membership (or shared self-categorization) between followers and leaders (as discussed here) and, on the other hand, the extent
to which a leader is seen to act in ways that champion the interests of the group he or she is part of (as discussed in the preceding
section). Leaders might be categorized as belonging to a specific group, and their followers may similarly define themselves
(or not define themselves) as part of that same group. Moreover, leaders are likely to vary in the extent to which they actively
champion the interests of their particular group (i.e., of a group that followers see themselves as belonging to or not). Indeed,
a leader may belong to a particular group, yet still be seen as doing little to stand up for the interests of that group and instead
be seen as championing the interests of other groups (that he or she may be a member of or not) or as championing his or her
personal interests. As a result, followers’ perceptions of leader authenticity and followership as a result of the leader’s
championing of collective interests are likely to depend on their perceptions of shared group membership with the leader, but
such judgments will be particularly pronounced when the leader champions the interests of a valued group — a group that followers
believe in and see as an important part of who they are.
Indeed, although a leader may be committed to principles that promote the pursuit of truth and may act consistently in ways
that are guided by his or her moral principles, there is rarely consensus among all potential followers about what “truth” is or
about what ends or whose interests particular moral principles should be serving. Along these lines, even Mahatma Gandhi’s leadership
— which centered on fervent commitment to key generic values and beliefs (e.g., a belief in non-violent protest) and led to
him being seen by many as a saint — was not seen as equally authentic by all. For example, some Hindu subgroups criticized his
national leadership for not being sufficiently pro-Hindu and other religious factions in India (in particular, Jews and Muslims)
were critical of his stance because it failed to serve their interests (Nanda & Nanda, 1985). Accordingly, it seems likely that shared
social self-categorization acts as a boundary condition for the relationships proposed in H1 and H2 such that support for these
hypotheses will be more pronounced under conditions of shared identity between leaders and followers. Thus we hypothesize:
H3. Shared social self-categorization will moderate the relationship between leaders’ championing of collective interests and authentic
leadership. Specifically, the relationship between leaders’ championing of collective interests and authentic leadership will
be more pronounced when would-be followers see a leader to be affiliated to a group that they see as self-defining (an ingroup)
rather than a different group (an outgroup).
H4. Shared social self-categorization will moderate the relationship between leaders’ championing of collective interests and followers’
followership (i.e., voting intentions). Specifically, the relationship between leaders’ championing of collective interests and
followership (i.e., voting intentions) will be more pronounced when would-be followers see a leader to be affiliated to a group
that they see as self-defining (an ingroup) rather than a different group (an outgroup).
The present research
In the present research, we conducted two studies in different contexts and using different methodologies to examine the impact
of followers’ perceptions of leaders’ championing of collective interests on authentic leadership. To test the causal impact of
the proposed relationships, in Study 1 we provide an experimental test by manipulating followers’ perceptions that leaders are
champions of collective (rather than personal) interests before assessing authentic leadership and followership (thereby testing
H1 and H2). In Study 2, our goal was to replicate this study in the field and also to examine the degree to which shared selfcategorization
between perceivers and leaders serves as a boundary condition for these hypotheses (thereby also testing H3
and H4).
Study 1
Method
Participants
A sample of 74 participants took part in this study. They were recruited on campus at The University of Queensland (in Brisbane,
Australia) and participated voluntarily. Missing data in one case resulted in a final sample of 73 participants. Participants’
mean age was 30.03 years (SD = 11.80) and 54 (73.0%) participants were female (one participant did not specify their gender).
Participants were randomly assigned to one of two experimental conditions (leader championing personal versus collective
interests).
Procedure and measures
The experiment was conducted one month after the 2013 leadership challenge to Australia’s Prime Minister (a reassessment of
the Prime Minister by all Members of her party) and prior to the 2013 Australian federal election. Participants were invited to
take part in a study entitled “Perceptions about contemporary leaders” and in both conditions they were presented with a
one-page commentary newspaper article about Bill Shorten, a senior Minister of the Australian Government at the time. The article
was identical in both conditions and discussed Shorten’s change of mind in switching his support from the then Prime

Minister in office (Julia Gillard) to the newly proposed Prime Minister (Kevin Rudd). The article discussed the fact that Bill Shorten
showed inconsistent behavior by first transferring his support from Rudd to Gillard in 2010 and then transferring it back from
Gillard to Rudd three years later.
The articles in the two conditions differed in emphasizing that the leader’s inconsistent behavior was motivated by the leader’s
advancement of either personal interests or collective interests. The article in the “personal interests” condition was titled “Bill
Shorten changes his mind to advance personal interests”. The article further emphasized that although he said that he was acting
to advance national interests, he changed his mind because he was pursuing his personal ambitions and that he regarded himself
as a suitable future Prime Minister. In contrast, in the “collective interests” condition the article was titled “Bill Shorten changes
his mind to advance collective interests”. The article then emphasized that Shorten had changed his mind because he was mindful
of Australia’s needs and that he believed changing the Prime Minister would be in the best interest of the nation.
Following the manipulation, participants in both conditions watched a four-minute video of a press conference in which Shorten
declared that he had changed his mind to support a new future Prime Minister (Kevin Rudd), and in which he also asked the
public to vote for this candidate in the upcoming federal election. Afterwards, participants in both conditions responded to the
measures below and then provided demographic data before being debriefed.
Manipulation check. Participants responded to three items that assessed the extent to which they believed that the leader had
acted to advance collective, rather than personal, interests (α = 0.82): “Bill Shorten changed his mind in order to advance the
interests of Australians as a whole”; “Bill Shorten changed his mind in order to advance his personal interests” [reversed scored]
on scales ranging from 1 (strongly disagree) to 7 (strongly agree); and “Bill Shorten changed his mind to advance…1 (personal interests)
to 7 (collective interests)”.
Authentic leadership. Participants responded on 5-point Likert scales ranging from 1 (strongly disagree) to 5 (strongly agree) to the
14 items (α = 0.90) of Neider and Schriesheim’s (2011) Authentic Leadership Inventory (ALI). This global authentic leadership
measure comprises four sub-dimensions: self-awareness (α = 0.69; three items: e.g., “Bill Shorten shows that he understands
his strengths and weaknesses”), relational transparency (α = 0.76; three items: e.g., “Bill Shorten clearly states what he
means”), internalized moral perspective (α = 0.77; four items: e.g., “Bill Shorten shows consistency between his beliefs and actions”),
and balanced processing (α = 0.60; four items: e.g., “Bill Shorten carefully listens to alternative perspectives before
reaching a conclusion”).
Followership. Participants responded to one item assessing the extent to which they were willing to follow Shorten’s appeal to vote
for the candidate he was proposing (“If elections were held tomorrow, which leader would you vote for?”). Participants were invited
to choose between the candidate Shorten was proposing (Kevin Rudd) and the alternative candidates from the main competing
parties (Tony Abbott, Christine Milne, or “Other”). To provide a test of H2, we combined the choices of alternative
candidates to assess whether participants were willing to vote for the candidate proposed by the leader or for an alternative
candidate.
Results
Preliminary analyses examining factors of authentic leadership
Means, standard deviations, and intercorrelations between variables are presented in Table 1. We followed the general recommendations
by Neider and Schriesheim (2011) to assess in each sample whether it is more appropriate to treat the four ALI dimensions
separately or together as part of an overall authentic leadership factor in the main analyses. We note that addressing
the question of whether or not a global or a component conceptualization of authentic leadership is more suitable is not a key
aim of the present paper; more definite tests and carefully designed studies with the aim of examining the factor structure in
multiple studies and samples are needed to address this matter conclusively. Nonetheless, it was critical to determine which conceptualization
is more appropriate in the present case. Although we do not intend to draw inferences with regard to the factor

Table 1
Study 1: Means, standard deviations, and intercorrelations between variables.
Variable Mean SD 1 2 2.a 2.b 2.c 2.d 3

  1. Manipulation check [collective interests] 3.67 1.37 –
  2. Authentic leadershipa 4.02 0.93⁎ 0.51⁎⁎ –
    2.a. Self-awareness 4.28 1.05 0.20 – –
    2.b. Relational transparency 4.08 1.25 0.61⁎⁎ – 0.64⁎⁎ –
    2.c. Internalized moral perspective 3.92 1.13 0.58⁎⁎ – 0.57⁎⁎ 0.69⁎⁎ –
    2.d. Balanced information processing 3.88 0.90 0.45⁎⁎ – 0.72⁎⁎ 0.71⁎⁎ 0.67⁎⁎ –
  3. Followership 0.45 0.50 0.15 0.17 0.20 0.15 0.20 0.04 –
    a Authentic leadership comprising its four sub-dimensions; Manipulation check was indicated on a 7-point Likert scale ranging from 1 (strongly disagree) to 7
    (strongly agree) and authentic leadership was indicated on a 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree); Followership was coded as 0
    (voting for candidate other than the one suggested by leader) and 1 (voting for candidate as suggested by leader).
    ⁎ p b 0.05.
    ⁎⁎ p b 0.01; N = 73.

Table 2
Study 1: Confirmatory-factor analyses results displaying fit statistics and model comparison statistics for higher-order model and alternative models of authentic
leadership.
Model χ2 df p χ2/df Δχ2 Δdf p RMSEA [90% CIs] CFI SRMR
Higher-order model 137.543 73 b0.001 1.884 – – – 0.111 [0.082, 0.139] 0.843 0.0813
Higher-order modela 143.594 74 b0.001 1.940 – – – 0.114 [0.086, 0.142] 0.831 0.0816
More parsimonious models
Orthogonal first-order model 305.540 77 b0.001 3.968 161.946 3 b0.001 0.203 [0.179, 0.227] 0.446 0.3355
Single-factor model 151.418 77 b0.001 1.966 7.824 3 0.049 0.116 [0.088, 0.143] 0.820 0.0832
Less parsimonious models
Oblique first-order model 133.744 71 b0.001 1.884 9.850 3 0.020 0.111 [0.082, 0.139] 0.848 0.0795
Bi-factor modelb 130.438 66 b0.001 1.976 13.156 11 0.250 0.116 [0.087, 0.146] 0.844 0.0768
N = 73; 95% CIs = 90% confidence intervals; All χ2 difference tests involve comparisons with the higher-order model with one error variance that was specified to
have a small positive value of 0.001.
a Because the error variance for the lower-order factor balanced processingwas estimated to be negative in the originalHOM, in the presentmodel the error variance
was specified to have a small positive value of 0.001.
b Because three error variances for the three items ALI 4, 9, and 10 were estimated to be negative in the original Bi-Factor model, in the present model the error
variances for these three items were specified to have a small positive value of 0.001.

structure of authentic leadership more generally as assessed by the ALI, providing a rigorous test of the evidence in relation to
higher-order factors contributes to the current debate in organizational psychology. Credé and Harms’ (2015) review of higherorder
factor treatment in leading organizational psychology journals revealed that in most studies researchers claimed to find support
for a higher-order factor, while on the basis of the Credé and Harms’ (2015) recommendations, interpretation of support for
a higher-order factor was not warranted in the majority of these cases.
To engage constructively in this debate, we adopted Credé and Harms’ (2015) recent comprehensive guidelines for testing and
interpreting evidence in support of (versus against) the appropriateness of higher-order models. A summary of the CFA results for
competing models is presented in Table 2 (see Appendix A for more detailed results). Overall, analysis of five sets of evidence for
(versus against) a higher-order model provided more support for than against a higher-order model of authentic leadership. On
the basis of these results, and consistent with our hypotheses and previous research that conceptualizes authentic leadership as a
global higher-order construct (Gardner et al., 2011), we therefore analyzed results at the higher-order level of authentic leadership.
We would note, though, that if the data are analyzed at the level of the sub-dimensions, results are largely identical.1
Main analyses
Manipulation check.Means as a function of experimental condition and inferential statistics are presented in Table 3. We conducted
a multivariate analysis of variance of experimental condition on the manipulation check and authentic leadership. Analysis revealed
an effect of experimental condition on the leader’s championing of collective interests, F(1.71) = 6.95, p = 0.010, MD =
0.82, 95%CIs [0.20, 1.44], Cohen’s d = 0.63, indicating that participants perceived a leader as a greater champion of collective interests
when that leader was said to be championing collective interests (M = 4.11, 95%CIs [3.66, 4.56], SD = 1.09) more than
when that leader was said to be championing personal interests (M = 3.29, 95%CIs [2.87, 3.71], SD = 1.49). This suggests that
our manipulation was successful.
Authentic leadership. Analysis revealed a significant effect of condition on authentic leadership, F(1.71) = 6.56, p = 0.013, MD =
0.54, 95%CIs [0.12, 0.96], Cohen’s d = 0.60. Participants saw a leader who championed collective interests as displaying greater
authentic leadership (M = 4.31, 95%CIs [4.00, 4.62], SD = 0.87) than a leader who championed personal interests (M = 3.77,
95%CIs [3.49, 4.06], SD = 0.92).
Followership. A chi-square test indicated that participants were more likely to follow the leader when the leader was championing
collective, rather than personal, interests, χ2 (1, N = 73)=4.77, p = 0.029. When the leader was championing personal interests,
33.3% of participants (13 out of 39) were likely to follow the leader’s proposal by voting for the candidate that the leader had
advocated. In contrast, 58.8% of participants (20 out of 34) followed the leader in this way when the leader was said to be
championing collective interests. An odds ratio of 2.86 indicates that followers were more than twice as likely to follow the leader
by voting for the candidate the leader endorsed when the leader was seen to be championing collective, rather than personal, interests.

If results are analyzed at the level of lower-order factors industry, self-awareness, relational transparency, internalizedmoral perspective, and balanced processing,
the pattern of results is largely identical. Condition had a marginally significant effect on self-awareness, F(1.71)= 3.74, p= 0.057, MD= 0.47, 95%CIs [−0.01, 0.95],
and significant effects on relational transparency, F(1.71)= 5.63, p= 0.020,MD= 0.67, 95%CIs [0.11, 1.24], internalizedmoral perspective, F(1.71)= 5.18, p= 0.026,
MD= 0.59, 95%CIs [0.07, 1.10], and balanced processing, F(1.71) = 4.56, p= 0.036, MD= 0.44, 95%CIs [0.03, 0.85].

Table 3
Study 1: Means (standard deviations in parenthesis), inferential statistics, and effect sizes for authentic leadership and followership as a function of experimental
condition.
Measure Condition: leader championing of Statistics and effect size
Personal interests
(n = 39)
Collective interests
(n = 34)
F(1.71) η2 Cohen’s d
Manipulation check 3.29 (1.49) 4.11 (1.09) 6.95⁎⁎ 0.01 0.63
Authentic leadership 3.77 (0.92) 4.31 (0.87) 6.56⁎ 0.01 0.60
Followership 33.33% (13 of 39) 58.82% (20 of 34) χ2 (1, N = 73) = 4.77
Manipulation check was indicated on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree) and authentic leadership was indicated on a
Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Followership was coded as 0 (voting for candidate other than the one suggested by leader)
and 1 (voting for candidate as suggested by leader).
⁎ p b 0.05.
⁎⁎ p b 0.01.

Discussion
Study 1 aimed to provide a causal examination of the extent to which perceptions of authentic leadership and acts of followership
are structured by a leader’s championing of collective, rather than personal, interests. Supporting H1, respondents regarded
a leader who championed collective interests to be a more authentic leader than a leader who championed personal interests.
Moreover, supporting H2, participants were more willing to follow a leader (by voting for the candidate that the leader had proposed)
when the leader was said (and seen) to be advancing collective, rather than personal, interests. Both effects were robust
and of moderate to large size (Cumming, 2014). In this way, experimental findings provide solid support for a causal model in
which a leader’s advancement of collective interests that are associated with a shared group membership serves to structure
both perceptions of their authentic leadership and group members’ acts of followership.
Yet despite this support for our main hypotheses, the inferences we can draw on the basis of the present findings are limited
in certain ways. First, the study only examined leaders who advanced either collective interests or personal interests. We manipulated
the variable of interest in this way largely because in many contexts, perceptions of a leader’s advancement of personal and
collective interests are likely to be negatively correlated. Nevertheless, as self-categorization theory suggests, it is possible for the
advancement of collective interests to be independent of advancement of personal interests (such that advancement of collective
interests is sometimes seen as advancement of personal interests and sometimes not). Furthermore, it is also possible that a leader
may fail to advance the collective interests of a particular (in)group by advancing the collective interests of an alternative
(out)group. Accordingly, to examine the broader impact of a leader’s advancement of collective interests, Study 2 aimed to explore
the impact of a leader being seen to be championing the collective interests of a given group independently of whether participants
regarded him or her to be simultaneously championing personal interests or those of an outgroup.
A second limitation of this first study arises from the fact that people can self-categorize as members of various groups — not
only at an abstract superordinate level (e.g., as an Australian) but also at less abstract lower levels (e.g., as a member of a particular
political party, e.g., the Labor party in the Australian context). It is thus conceivable that the impact of leaders’ advancement
of the collective interests of a lower-level identity (e.g., a particular department vis-à-vis other departments within an organization;
a political party vis-à-vis other parties within a country) is likely to depend on whether a would-be follower regards him- or
herself to be a member of that same lower-level group or of another alternative group (in ways envisaged by H3 and H4). To

Table 4
Study 2: Means, standard deviations, and intercorrelations between variables.
Variable Mean SD 1 2 3 4 4.a 4.b 4.c 4.d 5

  1. Follower self-categorizationa 0.51 0.50 –
  2. Leader group membershipa 0.50 0.50 −0.04 –
  3. Leader collective identity 4.70 1.86 0.09 −0.06 –
    advancement
  4. Authentic leadershipb 2.94 1.02 0.05 0.05 0.66⁎⁎ –
    4.a. Self-awareness 2.93 1.10 0.02 0.05 0.65⁎⁎ – –
    4.b. Relational transparency 2.86 1.16 0.05 0.13⁎ 0.61⁎⁎ – 0.86⁎⁎ –
    4.c. Internalized moral perspective 3.16 1.04 0.05 −0.04 0.64⁎⁎ – 0.82⁎⁎ 0.81⁎⁎ –
    4.d. Balanced information processing 2.78 1.05 0.05 0.06 0.60⁎⁎ – 0.90⁎⁎ 0.89⁎⁎ 0.81⁎⁎ –
  5. Followership 3.62 2.63 −0.03 −0.05 0.59⁎⁎ 0.81⁎⁎ 0.76⁎⁎ 0.77⁎⁎ 0.72⁎⁎ 0.78⁎⁎ –
    a Follower self-categorization coded as 0 (Liberal self-identifier) and 1 (Labor self-identifier) and leader groupmembership coded as 0 (Liberal leader) and 1 (Labor
    leader) were combined into shared versus non-shared self-categorization for subsequent analyses.
    b Authentic leadership comprising its four sub-dimensions; Leader collective identity advancement and followership (voting intentions) items were indicated on
    Likert scales ranging from1 (strongly disagree) to7(strongly agree); Authentic leadership itemswere indicated on scales ranging from1 (strongly disagree) to 5(strongly
    agree).
    ⁎ p b 0.05.
    ⁎⁎ p b 0.001.

Table 5
Study 2: Confirmatory-factor analyses results displaying fit statistics and model comparison statistics for higher-order model and alternative models of authentic
leadership.
Model χ2 df p χ2/df Δχ2 Δdf p RMSEA [90% CIs] CFI SRMR
Higher-order model 309.152 73 b0.001 4.235 – – – 0.113 [0.100, 0.126] 0.935 0.0375
Higher-order modela 309.735 74 b0.001 4.186 – – – 0.112 [0.099, 0.125] 0.935 0.0376
More parsimonious models
Orthogonal first-order model 1488.461 77 b0.001 19.331 1178.726 3 b0.001 0.269 [0.257, 0.281] 0.610 0.5630
Single-factor model 345.856 77 b0.001 4.492 36.121 3 b0.001 0.117 [0.105, 0.130] 0.926 0.0422
Less parsimonious models
Oblique first-order model 303.152 71 b0.001 4.270 6.583 3 0.086 0.113 [0.100, 0.127] 0.936 0.0371
Bi-factor modelb 206.784 64 b0.001 3.231 102.951 10 b0.001 0.094 [0.080, 0.108] 0.961 0.0263
N = 73; 95% CIs = 90% confidence intervals.
a Because the error variance for the lower-order factor self-awarenesswas estimated to be negative in the original HOM, in the presentmodel the error variancewas
specified to have a small positive value of 0.001.
b Because the error variance for ALI item2was estimated to be negative in the original bi-factormodel, in the presentmodel the error variancewas specified to have a
small positive value of 0.001. All χ2 difference tests involve comparisonswith the higher-ordermodelwith one error variance thatwas specified to have a small positive
value of 0.001.

address these issues, we examined whether or not these patterns were particularly clear among followers who self-categorize
themselves as a member of the specific group under the leader’s charge.

Study 2
In Study 2 we sought to replicate Study 1 and extend it in two important ways. First, we aimed to provide a more general test
of our hypotheses by examining the relationship between the championing of collective interests and authentic leadership in the
field. Second, we set out to quantify the degree to which the relationship between a leader’s championing of collective (group)
interests and authentic leadership perceptions and followership are more pronounced for perceivers who self-categorize as members
of the group that a leader is leading (rather than as members of a different group). This involved a study with a geographically
and demographically representative sample of Australians two weeks prior to the country’s 2013 general election.
Method
Participants
Two hundred and fifty-five Australian citizens (118 males; 136 females; 1 did not specify) participated in the present study.
They were recruited from the general population via a professional online recruiting company (ORU) — Australia’s leading online
data sampling company that has access to the largest research panel in Australia and ensures representative sampling of the population
— and received $10 for their participation. Participants ranged in age from 18 to 83 years (M = 46.95; SD = 17.97) and
were geographically representative of the country’s electorate, being dispersed across Federal States largely consistent with population
density, residing in New South Wales (n = 81), Victoria (n = 65), Queensland (n = 50), South Australia (n = 22), Western
Australia (n = 20), Tasmania (n = 10), Australian Capital Territory (n = 6), and Northern Territory (n = 1).
Design and procedure
We assessed people’s perceptions of the authentic leadership of the two key contestants for the office of Prime Minister (Tony
Abbott and Kevin Rudd) among people who identified with one of the two major political parties in Australia (the Liberal and
Labor parties). In this way, the study allowed us to investigate people’s perceptions of authentic leadership in the context of a
momentous decision about the future of their country as well as to assess parallel effects on the degree to which they were prepared
to follow a leader by giving them their vote in the election. Stated more formally, the study employed a quasi-experimental
2 (follower self-categorization: Liberal versus Labor) × 2 (leader group affiliation: leader of the Liberal party [Tony Abbott] versus
leader of the Labor party [Kevin Rudd]) between-participants design, in which leaders’ championing of the collective interests of
their party (hereafter referred to as collective identity advancement) was co-measured.
Participants were invited to take part in a study entitled “Survey about evaluation of various political leaders”. They then were
asked to indicate the political party that they identified most strongly with based on the single-item social identification measure
(SISI; Postmes, Haslam, & Jans, 2013): “If you had to choose between the different major political parties (e.g., Liberal party,
Labor party), which do you identify most strongly with? I identify most strongly with…Liberal party/ Labor party/ Greens/
Other”. If respondents self-categorized in terms of a party other than one of the two major parties, the survey was terminated
and they were thanked for their participation. If they self-categorized as either Liberal or Labor, they then responded to a series
of items assessing the leadership of either Tony Abbott (leader of the Liberal Party) or Kevin Rudd (leader of the Labor party).
More specifically, participants responded on 7-point Likert scales ranging from 1 (strongly disagree) to 7 (strongly agree) to four
items (α = 0.94) assessing leader collective identity advancement from the Identity Leadership Inventory (ILI; Steffens et al.,
2014b): “[This leader] promotes the interests of [the Liberal/Labor Party]”, “This leader acts as a champion for [the Liberal/

Table 6
Study 2: Hierarchical regression analyses for authentic leadership and followership as a function of leader collective identity advancement and shared self-categorization between follower and leader.


Variable Authentic
leadership
Followership (voting intentions)
Step 1 Step 2 Step 1 Step 2
β b [95% CIs] SE t β b [95% CIs] SE t β b [95% CIs] SE t β b [95% CIs] SE t
Main effects
Shared self-categorization 0.57 1.15 [0.99, 1.32] 0.09 13.61⁎⁎ 0.53 1.07 [0.89, 1.24] 0.09 12.04⁎⁎ 0.82 4.31 [4.00.4.62] 0.16 27.46⁎⁎ 0.75 3.94 [3.65, 4.24] 0.15 26.19⁎⁎
Leader collective identity
advancement
0.37 0.38 [0.30, 0.46] 0.04 8.98⁎⁎ 0.31 0.32 [0.22, 0.41] 0.05 6.62⁎⁎ 0.16 0.43 [0.28, 0.59] 0.08 5.51⁎⁎ 0.05 0.14 [−0.02, 0.30] 0.08 1.72
Two-way interaction
Shared self-categorization x Leader
collective identity advancement
0.14 0.29 [0.09, 0.48] 0.10 2.87⁎⁎ 0.23 1.26 [0.93, 1.59] 0.17 7.46⁎⁎
Model
df 252⁎ 251 251 250
ΔR2 0.68⁎⁎ 0.01⁎⁎ 0.83⁎⁎ 0.04⁎⁎
R2 0.68⁎⁎ 0.69⁎⁎ 0.83⁎⁎ 0.87⁎⁎
Standardized coefficients are reported; Due to missing data, degrees of freedom vary.
⁎ p b 0.05.
⁎⁎ p b 0.01; N = 255.