When controversial leaders with charisma are effective: The
influence of terror on the need for vision and impact
of mixed attitudinal messages
ERNESTINE H. GORDIJN1*
AND DIEDERIK A. STAPEL2
1
Department of Social and Organizational
Psychology, University of Groningen, The
Netherlands
2
Tilburg Institute for Behavioral Economics
Research (TIBER), Tilburg University, The
Netherlands
Abstract
We investigated the idea that a charismatic leader with a controversial message is most likely to
persuade people in times of terror, because in those times people have a high need for vision, and vision
is what a charismatic leader provides. In addition, we argued that the leader's message should contain
a pro-attitudinal position as well, as this makes the counter-attitudinal message more palatable. In line
with our hypotheses, we found in Experiment 1 that thinking about terrorism increases people's need
for vision. Experiment 2 revealed that only when people have a high need for vision they will be
influenced by a controversial charismatic leader. Experiment 3 showed that existential threats also
directly increase the influence of a controversial charismatic leader. Further, this was especially so
when the charismatic leader was both attractive and communicated his message in a charismatic way.
Finally, Experiment 4 revealed that after thinking about their own death or about terrorist attacks,
people were most likely to be persuaded by a controversial charismatic leader whose counterattitudinal
message also contained pro-attitudinal statements. Together, this research suggests that in
times of terror people's need for vision increases, which opens them up to a counter-attitudinal message
of a charismatic leader as long as this message also includes some pro-attitudinal statements.
Copyright # 2006 John Wiley & Sons, Ltd.
In this paper we investigate under what circumstances charismatic leaders with controversial ideas are
most likely to be effective in changing the attitudes and behaviors of their audience. We argue that in
times of terror, people have a high need for vision which opens them up to the influence of a charismatic
European Journal of Social Psychology
Eur. J. Soc. Psychol. 38, 389­411 (2008)
Published online 22 November 2006 in Wiley InterScience
(www.interscience.wiley.com) DOI: 10.1002/ejsp.411
*Correspondence to: Dr Ernestine H. Gordijn, Department of Social and Organizational Psychology, University of Groningen,
Grote Kruisstraat 2/1, 9712 TS Groningen, The Netherlands. E-mail: e.h.gordijn@rug.nl
Copyright # 2006 John Wiley & Sons, Ltd.
Received 14 February 2006
Accepted 4 September 2006
leader. However, we further argue that under such circumstances, persuasion is most likely when the
controversial message the charismatic leader communicates also includes pro-attitudinal statements.
These hypotheses were inspired by observations about the rise of Pim Fortuyn, a controversial
charismatic political leader in The Netherlands. The rise of this popular political leader--who attracted
an amazingly large following in a very short time--not only raised the question of `what makes
charisma work,' but also pointed to possible answers. Before relating these insights to theorizing with
respect to charisma, management of terror, and persuasion, and introducing the four experiments that
tested our model (see Figure 1), we describe the observations that sparked our thinking.
PIM FORTUYN: A CASE STUDY IN EFFECTIVE CHARISMATIC LEADERSHIP
Pim Fortuyn, who was murdered on 6 May 2002, was the leader of a new and controversial political
party in The Netherlands. Just before his assassination, a survey revealed that Fortuyn's party, which
was founded only a few months earlier, would be the largest party in the Dutch parliament,1
making
Fortuyn the presumptive Prime Minister of The Netherlands. This idea came as a shock to many, given
that he had controversial political views according to Dutch society at the time. Why was Fortuyn so
effective in persuading such a large number of people? Most of the explanations given were focused on
his personality, his appearance, and the way he communicated his message, or in other words on his
charisma (see e.g., Baars, Etman, Luitwieler, & Visser, 2002; Ellemers, 2002). Fortuyn behaved in
ways that are typically associated with charisma: he excelled at making difficult things (look) simple:
he could explain complicated ideas quite effectively and with seeming ease. Put another way, he was
able to take complicated political issues and explain them in a clear and concise manner. Fortuyn was
physically attractive and always looked directly into the camera. Fortuyn was a good debater and he
presented a vision: It always seemed clear what he wanted and how he wanted to get there. All these are
components of a charismatic style (e.g., Kirkpatrick & Locke, 1996).
But is charismatic style sufficient to explain Fortuyn's rise to political influence? We think not and
argue that Fortuyn was so influential because the time (right after 9/11/2001) in which he voiced his
controversial views made people very open to charismatic leadership. Although The Netherlands were
Terror
Salience
Need for
Vision
Charisma
(Mixed)
Message
Content
Message
Agreement
Figure 1. The terror-vision-charisma model
1
This quote was taken from the following website address: http://www.rnw.nl/achtergronden/html/fortuyn020506.html.
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
390 Ernestine H. Gordijn and Diederik A. Stapel
not directly affected by the terrorist attacks, fear of terrorism, and social crisis were widespread
(Dumont, Yzerbyt, Wigboldus, & Gordijn, 2003). The aftermath of the 9/11 terrorist attacks is likely to
have provided a charisma-conducive environment (see Ehrhart & Klein, 2001) in The Netherlands.
However, even though people may have been open to charismatic influence because of fear of
terrorism, why were they open to ideas that were at the time very controversial? Perhaps Fortuyn's
persuasive success was also the result of the fact that his political views were a strange mix of
conservative/progressive, left/right extremist views. He, for example, claimed to be against
immigration, but wanted to legalize all the illegal aliens already living in The Netherlands. Moreover,
he completely dismissed Islam and everything associated with it, but loved to tell stories about his
(sexual) affairs with young Arabic men. According to Kolbert (The New Yorker, 9 September 2002),
`Fortuyn's flamboyant gayness helped make his reactionary xenophobia publicly palatable' (p. 106).
We think that this quote highlights Fortuyn's effect on a more liberal audience. At the same time,
however, one could argue that people whose attitudes towards gays or immigrants were not very liberal
to begin with, may have started to think in a much more positive way about gays because of Fortuyn.
This analysis suggests that Fortuyn's charisma really worked well because his message was a new and
provocative mix of views that were typically associated with both left and right wing politics.
In sum then, this illustrative case study of the sudden success of Fortuyn led to the development of
the argument that he was influential not only because he was charismatic, but also because he
communicated a mixed ideological message in times of terror. This case study suggests two important
features that determine whether controversial charismatic leaders are likely to change people's
attitudes: a charisma-conducive environment, which creates a need for vision, and the communication
of a mixed ideological message. When these factors are combined, the impact of persuasive efforts will
be especially effective. In this research, we test this conjecture in four experiments. We will first
elaborate on the factors that we focused on in our research: charisma, a charisma-conducive
environment, and communication of a mixed ideological message.
THE ROLE OF CHARISMA
According to Kirkpatrick and Locke (1996), someone who has charisma articulates a clear vision and
shows how to implement the vision, all while using a charismatic communication style. A vision may
be effective because it touches upon listeners' most basic needs and values. The ability to articulate a
vision is essential for charismatic leadership. Visions are general and are never fully achieved in
practice. Yet, to be charismatic, a leader must go beyond simply communicating a vision. The leader
has to show how to implement the vision, clarify what is to be accomplished, and describe how the task
is to be done. In order to communicate both vision and vision implementation, Kirkpatrick and Locke
(1996) further argue that a charismatic communication style is needed. Such a communication style
entails, for example, speaking with a captivating voice, making direct eye contact, showing animated
facial expressions, being attractive, and having a dynamic interaction style (see also, e.g., Bass, 1985;
Conger & Karungo, 1987).
However, as we argued before, having a charismatic personality is probably not sufficient for
becoming a persuasive leader (e.g., Weber, 1947). According to Ehrhart and Klein (2001); see also, for
example, Bligh, Kohles, & Meindl, 2004; House, Spangler, & Woycke, 1991; Klein & House, 1995),
charisma resides in the relationship between a leader who has charismatic qualities and those of his
followers who are open to charisma, within a charisma-conducive environment.
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 391
A CHARISMA-CONDUCIVE ENVIRONMENT: THE AFTERMATH
OF THE 9/11 TERRORIST ATTACKS
When is an environment open to charisma? In what social situations is charisma most likely to flourish?
A condition of social crisis or rapid social change is the most frequently discussed precondition that
favors the emergence of charismatic influence (e.g., Becker, 1973; Beyer, 1999; Bord, 1975; Friedland,
1964; Haslam et al., 2001; House et al., 1991; Hunt, Boal, & Dodge, 1999; Pillai, 1996; Shamir &
Howell, 1999; Weber, 1947). We argue that acts of terrorism such as the 9/11 terrorist attacks might
result in such feelings of threat and social crisis. When facing terrorism, people look for ways to get out
of this crisis and for signs and ideas to make things go better in the future. In other words, people will
look for leaders with vision (see, e.g., Bligh et al., 2004). This is precisely the type of information a
charismatic leader provides. Charismatic leaders, like Pim Fortuyn, tell people how they should
proceed--they have visions and they tell their audience how to implement these visions. Such a
person's visionary message gives their audience exactly what they are looking for: clear answers about
how to perceive reality and about how to proceed.
Some recent research is consistent with this idea. Cohen, Solomon, Maxfield, Pyszczynski, and
Greenberg (2004) showed that when people thought about their own mortality, they were more likely to
prefer a political candidate with a charismatic leadership style than a leader with a task-oriented or
relationship-oriented leadership style. On the other hand, when people were not facing such an existential
threat the charismatic leader was the least preferred candidate. This result supports Terror Management
Theory (e.g., Greenberg, Solomon, & Pyszczynski, 1986; Pyszczynski, Greenberg, & Solomon, 1999).
According to this theory, people who are confronted with real terror, such as their own mortality, deal with
this existential threat by searching for socially shared conceptions of reality (a cultural worldview), and for
a positive self-image. In this way people try to reduce the strong feelings of uncertainty they experience
after being confronted with this existential threat (Dechesne & Kruglanski, 2004).
Thus, when people are facing existential threats, such as their own mortality, they are open to
information that tells them how to proceed. The question of course is whether thinking about your own
mortality has similar consequences as facing terrorism. Pyszczynski, Solomon, and Greenberg (2003) and
Dechesne and Kruglanski (2004) make a theoretical case that acts of terrorism, such as the 9/11 attacks
could have the same effects as a mortality salience manipulation (see also Skitka, Bauman, & Mullen,
2004). And indeed, Landau et al. (2004) found that subliminal messages of `9/11' or `WTC' activated
death-related thoughts. Landau et al. also showed that reminders of mortality or 9/11 increased support for
the incumbent presidential candidate George W. Bush and decreased support for his contender in the 2004
elections John Kerry. They argued that it is likely that Bush has certain elements of a charismatic style,
which might, in combination with mortality or terror salience, have led to increased support. However,
because they could not experimentally manipulate whether Bush has charisma or not, it is sill unclear
whether Bush was more popular after priming terrorism because of a charismatic style or because of other
reasons. Thus, it is still not clear whether fear of terrorism increases the attractiveness of a charismatic
leader. Moreover, it also is not clear whether such a charismatic leader can actually change people's
attitudes under these circumstances. In the present research, we argue people who have a high need for
vision due to salience of terror are most likely to be persuaded by a charismatic leader with a
counter-attitudinal point of view when this leader communicates a mixed attitudinal message.
PERSUASION THROUGH THE USE OF MIXED ATTITUDINAL MESSAGES
Our analysis suggests that Fortuyn might have been successful because he was charismatic in a
charisma-conducive environment (i.e., right after 9/11), in other words, in an environment that made
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
392 Ernestine H. Gordijn and Diederik A. Stapel
people look for a leader with vision. But that is not only what made him successful. How could he have
persuaded people to agree with his provocative, extremist, and to many, clearly counter-attitudinal
views? After all, when facing existential threats, it seems rather strange that people might be open to
ideas that are different from what they had previously believed. To be more specific, Greenberg et al.
(1990) showed that reminding people of their mortality increases attraction to those who consensually
validate their beliefs and decreases attraction to those who threaten their beliefs. Counter-attitudinal
ideas may challenge people's cultural worldviews, and thus increase their uncertainty. Moreover, a
messenger with a counter-attitudinal message is likely to be seen as someone from an out-group
(Gordijn, Postmes, & De Vries, 2001). As research has shown, in times of terror people are known not
to be open to out-groups. Instead, they prefer their in-group (see e.g., Baldwin & Wesley, 1996;
Castano, Yzerbyt, Paladino, & Sacchi, 2002; Florian & Mikulincer, 1998). This suggests that it is likely
that a messenger's message will be discarded when this message is counter-attitudinal. How can a
messenger avoid such a dismissal?
Inspired by the Fortuyn case, we believe that a good strategy to avoid being seen as the out-group is
to provide a `mixed message.' Such a mixed message includes ideas that people will agree with in
addition to ideas that they will not agree with. We believe that such a message can reach more people.
That is, they can agree with at least a part of the message. When a counter-attitudinal message includes
statements about issues with which one already agrees, it can become a message with which people can
identify. In this way, the message can be attributed to `people like us' or, in other words, the `in-group'.
When a message is attributed to an in-group, the counter-attitudinal part will be seen as more palatable,
or at least worth considering (see e.g., Mackie, Gastardo-Conoco, & Skelly, 1992; Mackie, Worth, &
Asuncion, 1990). This reasoning suggests that a mixed message could be a good way of reaching many
people, even in times of terror, when people are less open to those who are different, and more open to
those who are part of the in-group. Consequently, it may be important to sell the message in such a way.
Surprisingly, to date the influence of such mixed messages on persuasion has not been studied.
CURRENT RESEARCH
To sum up, our analysis of the Fortuyn case and its relation to social psychological theorizing, suggests
that Fortuyn might have become popular and influential because of his charisma, his mixed ideological
message, and the feelings of terror and social crisis after 9/11 that surrounded Fortuyn's rise to power,
which increased people's need for vision. This analysis suggests new and general insights with respect
to the persuasive impact of charismatic leaders. In this paper we tested a model (see Figure 1) according
to which a charismatic leader will be influential during times of terror because in times of terror
people's need for vision increases. Moreover, in line with the model we aimed to show that the
charismatic leader with a counter-attitudinal point of view will be especially effective in influencing
these `vision craving' people when his message is `mixeď; that is, when a leader combines his
counter-attitudinal message about a particular political issue with pro-attitudinal statements about
another political issue.
We test this model in four experiments. In the first two experiments we will show that existential
threats increase the effectiveness of a charismatic leader, because fear of terrorism increases people's
need for vision (Experiment 1), which makes it likely that they are influenced by a charismatic leader
with a controversial mixed message (Experiment 2). In the third experiment we further examined the
concept of charisma, by distinguishing the leader's attractiveness from visionary wording of the
leader's persuasive message and examine their independent influence on attitude change. Moreover, we
aimed to directly test whether existential threats indeed lead to influence of a charismatic leader that
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 393
communicates a controversial mixed message. In the final study we examined the differential influence
of mixed versus one-sided controversial messages, aiming to show that only charismatic leaders who
communicate mixed messages are influential when people experience existential threats, such as fear of
terrorist attacks or their own death. We also included a behavioral measure to examine whether not only
attitudes but also real behavior is influenced.
EXPERIMENT 1
As argued before, acts of terrorism such as the 9/11 terrorist attacks might result in feelings of
existential threat and social crisis. In such times people are likely to look for directions on how to
proceed. In other words, they are in need of vision. To examine this idea, in Experiment 1 we measured
need for vision after a terror threat. To show that thinking about a terror-related event rather than just
thinking about something negative, such as pain, increases the need for vision, in the present
experiment we included a control condition in which people have to think about being in pain (see, e.g.,
Arndt, Greenberg, Solomon, Pyszczynski, & Schimel, 1999) Moreover, we included a mood measure
to test whether our findings with respect to need for vision cannot be attributed to a change in mood
rather than to terror salience.
Method
Participants and Design
Thirty-nine undergraduates from the University of Groningen participated in the experiment, which
had two experimental conditions (Terror Salience: terrorist attack salience vs. control: pain salience).
Participants received partial course credit for their participation.
Procedure
Participants were asked to fill out a personality questionnaire. It was said that this questionnaire
consisted of different sections that measured different aspects of personality. In line with the cover
story, participants were first asked to answer questions about how they interact with other people. The
second section, which included the terror salience manipulation, was said to measure how participants
think about issues such as terror (in the terror salience condition) or about pain (in the pain salience
condition). With respect to the terror salience condition we reminded participants of the 9/11 terrorist
attacks in New York in 2001, and the 3/11 terrorist attacks in Madrid in 2004, and argued that it is clear
that the Netherlands could also become a target of attacks. Two open-ended questions were asked: (1)
`Describe briefly the emotions and thoughts that thinking about the possibility of terrorist attacks in the
Netherlands arouses in you; and (2) `Jot down, as specifically as possible, what would happen when
terrorist attacks such as on 9/11 in New York or on 3/11 in Madrid would happen in the Netherlands'
(see for a comparable manipulation Landau et al., 2004). In the control condition (pain salience) we
told participants that we are interested in how people think about pain (dental pain, headaches,
toothaches, etc). Two open-ended questions were asked: (1) `Describe briefly the emotions and
thoughts you have when you are in serious pain, such as dental pain or a bad headache; and (2)' Jot
down, as specifically as possible, what would happens to you when you are in such pain' (see, e.g.,
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
394 Ernestine H. Gordijn and Diederik A. Stapel
Arndt et al., 1999). After this, participants were asked to fill out the third section of the questionnaire
that contained the dependent measures. The first part included 10 statements about their current mood
(e.g., `I feel sad,' `I feel happy' (recoded), `I feel down'; Cronbach's alpha  .85; a higher score implies
a more negative mood). Moreover, the questionnaire included several irrelevant statements, as well as
two statements about their need for vision (i.e., `I need someone who can offer me solutions,' `I would
like to listen to someone who can tell me how to proceeď; r .78). Answers were given on five-point
Likert scales (1  absolutely disagree; 5  absolutely agree). Afterwards participants were debriefed
and thanked.
Results and Discussion
Mood
A one-way analysis of variance (ANOVA) on the mood measure revealed no differences as a function
of the manipulation, F(1, 38)  1.55, p < .22; M  2.05, SD  0.59. The absence of an effect on mood is
in line with all previous research that investigated terror management and suggests that the effects of
thinking terror-related thoughts are not mediated simply by general negative affect (see, e.g.,
Greenberg et al., 1990; Landau et al., 2004).
Need for Vision
A one-way ANOVA on the vision measure revealed a significant effect of the manipulation, F(1,
38)  5.47, p < .03, h2
 0.13. As predicted, participants reported a higher need for vision (M  3.68,
SD  1.57) in the terror salience condition than in the pain salience condition (M  2.61, SD  1.26).2
The findings from this first experiment support the hypothesis that salience of terror-related events
increases people's need for vision: when people think about terror-related events, in this case the
possibility of terrorist attacks, they report that they are in need of someone who can offer them
solutions, and that they would like to listen to someone who can tell them how to proceed. Moreover,
results revealed that salience of other negative events, such as dental pain or bad headaches, does not
increase this need for vision.
EXPERIMENT 2
In Experiment 1 we found evidence for the idea that facing existential threats increases people's need
for vision. In Experiment 2 we aimed to show, that such an increased need for vision is necessary for
influence of a charismatic leader by manipulating people's need for vision and exposing them to either a
charismatic or non-charismatic leader that provides a mixed message.
2
When controlling for mood differences, the effect of the manipulation remains significant on need for vision, suggesting that
mood was not responsible for the differences in need for vision.
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 395
Method
Participants and Design
In this experiment 96 psychology freshmen from the University of Groningen participated in exchange
for course credit. The design was a 2 (Need for Vision: low vs. high) Â 2 (Charisma: low vs. high)
factorial design.
Procedure
Participants were invited to the laboratory where they were seated in front of PC's in cubicles. All
instructions, experimental information, and questions appeared on the screen. Answers were given via
the keyboard. Participants were told that they were participating in several studies. The `first study,' an
alleged personality test, was our manipulation of need for vision. The participants were told that this
test was based on Belbin's (1981) team role test, which is a test that is used to measure the roles people
have in teams. It was said that a new test has been developed by Harvard University which was more
general and could be used for personnel selection, but also for selecting students. Participants were told
that research in the USA had revealed that the test was predictive of being a successful student and
useful for companies to select new employees. It was then argued that in our research we wanted to test
the validity of a Dutch translation of this test. Participants were told that they would receive a
personality profile after filling out the test. The alleged test (which actually was modeled after Belbin's
team role test, Belbin, 1981) included 44 statements that appeared to test people's behavior with respect
to other people and with respect to their job. After filling out the questionnaire, the computer calculated
their so-called personality score. They were shown a profile that included a table in which they could
see how they scored with respect to the different personality styles (i.e., `integrator,' `secretary,'
`specialist,' `criticizer,' and `visionary'). Participants in the high need for vision condition were told
that their answers mostly matched the profile of the integrator, and least that of the visionary. They
received a description of what an integrator is, what role an integrator has, which behavior an integrator
should avoid, and with which personality type integrators should try to work with. It was said that an
integrator is someone who is sensitive, friendly, and supportive; someone who is very important for
keeping the group together. Moreover, an integrator is someone who is good at organizing and carrying
out ideas. It was further said that an integrator should avoid not being open to new and different ideas,
and that especially visionary people complement their strengths. So they should look for people with
vision.
In the low need for vision condition, participants were told the opposite. So they were told that their
answers mostly matched the profile of the visionary, and least that of the integrator. They are told that a
visionary person is someone with new, and very original ideas, and is someone who has the best ideas
for the future, and comes up with new solutions for old problems. They are further told that visionary
people are the ones who inspire others. It was further said that a visionary person should avoid being
arrogant to the people who they need to implement their ideas, and that they should look for people who
are integrators to complement their strengths. It should be noted that in both conditions, participants
received equal amounts of positive and negative feedback in order not to influence self-esteem in a
different way.
After this participants received a questionnaire with statements to check whether the manipulation
worked. They were asked to what extent they agreed with the statements that they were in need for
people with vision, that they wanted to be open to people with new ideas, and that they felt the need for
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
396 Ernestine H. Gordijn and Diederik A. Stapel
people who offered solutions for difficult problems (Cronbach's alpha  .603
). Participants responded
on seven-point Likert scales (1  absolutely not; 7  absolutely).
After this participants received another relatively long questionnaire that was designed to
manipulate charisma. In this questionnaire, which was titled `Students and politics,' participants were
presented with a paper allegedly taken from a national magazine. The title of this paper was `student
leader pleads for an increase in study load,' which is an opinion with which most students disagree. The
paper stated that `it is long known that only a small number of students are able to finish their degree in
4 years. This would suggest that study load is too high. However, the recently founded student
organization `Studenten Belangen Nederlanď (SBN: Student Interests Netherlands), actually pleads
for an increase in study load.' It was further stated that `last week, the leader of the SBN argued for this
deviant point of view during a meeting.'
Charisma was manipulated as follows: in the charismatic condition, the student leader was
described as powerful, self-assured, and attractive. Furthermore, parts of his speech were quoted in the
paper, which demonstrated his charismatic way of arguing. That is, his use of language revealed vision
(e.g., `What do we want? We want an education at the university that means more than it does now'),
and vision implementation (e.g., `How should this be done? Thaťs simple, we should have smaller
classes in which students can be encouraged to perform welľ). Finally, the paper included a picture of
the student leader: an attractive young man, who directly faced the camera. In the non-charismatic
condition, the student leader's position, which was of course similar in content to the charismatic
condition, was presented in a non-charismatic way. However, no reference was made to `what we want,
or how we should do it.' The arguments were just stated directly (e.g., `An education at the university
should mean more than it does now'; There should be smaller classes in which students can be
encouraged to perform welľ). Moreover, the picture of the student leader showed a profile of young
man who was not attractive. As a manipulation check participants were asked at the end of the
questionnaire to what extent they thought that the leader was charismatic (1  absolutely not;
9  absolutely).
In both conditions, the student leader provided a mixed message. In addition to arguing for an
increase in study load (focal counter-attitudinal issue, i.e., the attitude that we intended to change), he
also called for more involvement of students in the university administration (pro-attitudinal issue) and
an increase in grants (pro-attitudinal issue).
After reading the paper, we measured the measured attitudes towards the focal issue (i.e., `Study
load should be increaseď) and the two other issues (i.e., `Students should be more involved in
university administration'; `Grants should be increaseď). Participants responded on a nine-point Likert
scales (1  absolutely not; 9  absolutely). These items were embedded in questionnaire that contained
other irrelevant items with respect to student life. Afterwards participants were debriefed and thanked.
Results and Discussion
Need for Vision
The 2 (need for vision) Â 2 (charisma) ANOVA on the measure of need for vision only revealed a main
effect for need for vision, F(1, 92)  6.24, p < .02, h2
 0.06. As intended, results show that participants
in the high need for vision condition had a higher need for vision (M  5.73, SD  0.77) than
participants in the low need for vision condition (M  5.33, SD  0.84).
3
We realize this alpha is quite low. However, when analyzing the findings with respect to each item separately, a similar pattern is
found, and we therefore decided to group them into a scale.
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 397
Attitudes
As intended, the attitude towards the increase in study load, which is the counter-attitudinal focal issue
in this experiment, did not significantly correlate with either the attitude towards the increase in
involvement in university administration (r  .14, ns), or with the attitude towards an increase in the
grant they receive (r  .19, ns), while these last two did correlate with each other (r .36, p < .01).
We first analyzed the attitude towards the focal issue. The 2 (need for vision) Â 2 (charisma)
ANOVA on the focal attitude revealed a trend for a main effect for need for vision, F(1, 92)  3.52,
p < .07, h2
 0.04, as well as a significant interaction effect, F(1, 92)  5.93, p < .02, h2
 0.06. Simple
effects analyses showed that need for vision had no effect within the low charisma condition, F(1,
92)  0.14, ns, but it did have an effect within the high charisma condition, F(1, 92)  8.85, p < .01,
indicating that people who were confronted with a charismatic leader were less open to him when they
were in a low need for vision rather than high need for vision (See Figure 2).
We combined the two other attitude measures as they significantly correlated. The 2 (need for
vision) Â 2 (charisma) ANOVA on this combined attitude measure did not reveal any significant effects
(all F's < 1.55, all p's > .22).
The findings from this second experiment support the hypothesis that a charismatic leadership style
is only effective in influencing people's attitudes when they have a high need for vision. We argued that
a need for vision follows from existential threat, such as the possibility of terrorist attacks or thinking
about one's own mortality (see Experiment 1 for the influence of thinking about terrorist attacks). In the
next experiment we aimed to show directly that a charismatic leader is most effective when people
think about terror-related events. In this study we increased existential threat by having people think
about their own mortality.
EXPERIMENT 3
In Experiment 3, our main aim was to show that a charismatic leader who communicates a controversial
mixed message is most effective when people think about terror-related events, such as their own
mortality. Moreover, another aim of Experiment 3 was to deal with some confounds in our charisma
manipulation. That is, in the charismatic condition we not only manipulated the charismatic wording of
the message, but we also presented participants with the picture of an attractive leader. We combined
Figure 2. Attitude towards an increase in study load as a function of charisma and need for vision
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
398 Ernestine H. Gordijn and Diederik A. Stapel
these features, because we believe that both aspects can be part of the concept of charisma. As
Kirkpatrick and Locke (1996) said, `charisma is a complicated and multidimensional concept.' It might
be possible, however, that only one of these factors--either the picture of the attractive leader or the
charismatic wording of the message--lead to our findings rather than the combination of factors. We
therefore manipulated attractiveness and charisma independently. Furthermore, we added a charisma
manipulation check and a baseline condition in which no information was presented as these were
missing in Experiment 2.
Method
Participants and Design
In this experiment 102 undergraduates from the University of Tilburg participated during a mass
psychological testing session. The design was a 2 (Terror Salience: mortality salience vs. control: TV
salience) Â 3 (Charisma: picture only vs. vision only vs. picture vision) factorial design. We also
included a baseline condition (n  15) that received no information.
Procedure
The first questionnaire of this experiment concerned the manipulation of terror salience. In this
questionnaire participants received a mortality salience manipulation or received a TV salience
manipulation (the control condition). In the mortality salience condition the title of the questionnaire
was `Thinking about your own death.' In this questionnaire two open-ended questions were asked:
(1) `Describe briefly the emotions and thoughts that your own death arouses in you'; and (2) `Jot down,
as specifically as possible, what you think will happen to you when you die and once you are deaď (see
e.g., Rosenblatt, Greenberg, Solomon, Pyszczynski, & Lyon, 1989). In the control condition
participants received a questionnaire about watching television. In this questionnaire two open-ended
questions were asked: (1) `Describe briefly the emotions and thoughts that thinking about watching
television arouses in you'; and (2) `Jot down, as specifically as possible, what you recently watched on
television' (see for a similar control condition Dechesne, Greenberg, Arndt, & Schimel, 2000).
After this questionnaire participants received another questionnaire that was designed to provide a
counter-attitudinal message from the SBN (e.g., study load should be increased; see Experiment 2) and
to manipulate charisma. In all conditions (except of course for the baseline condition, which did not
receive any information, but only had to fill out the questionnaire) participants were presented with the
same arguments. Furthermore, in all conditions, the message was mixed. So, participants heard that the
SBN was in favor of an increase in study load (the focal counter-attitudinal issue), an increase in grants
(pro-attitudinal issue), and an increase in student involvement in university administration
(pro-attitudinal issue). The picture vision condition was similar to the charisma condition from
Experiment 2. In the picture only condition the participants saw the picture of the attractive student
leader with the arguments, but the arguments did not include vision or vision implementation
statements (e.g., `what do we want, we want . . .'; `how should this be done'). In the vision only
condition they received the same text as in the vision  picture condition, but the participants did not
see a picture of the leader. The dependent variables were the same as in Experiment 2. As a
manipulation check participants were asked at the end of the questionnaire to what extent they thought
that the leader was charismatic (1  absolutely not; 9  absolutely). We also included a measure about
the extent to which they thought the leader was offering solutions to the problems to examine their
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 399
effects of charisma. In the baseline condition only attitudes were measured. Afterwards participants
were debriefed and thanked.
Results and Discussion
Charisma
The 2 (terror salience) Â 3 (charisma) ANOVA on the measure of charisma only revealed a main effect
for charisma, F(2, 81)  3.03, p  .05, h2
 0.07. Results revealed that participants in both the vision
condition and the vision  picture condition judged the student leader to be more charismatic
(M  5.50, SD  1.53 (F(1, 81)  4.49, p < .05), and M  5.53, SD  1.25 (F(1, 81)  4.84, p < .05),
respectively), than participants in the picture condition (M  4.63, SD  1.84).
Solutions Offered
The 2 (terror salience) Â 3 (charisma) ANOVA on the measure of the extent to which the leader
provided solutions revealed a main effect for charisma, F(2, 81)  3.81, p < .05, h2
 .09. Results
revealed that participants in the vision  picture condition judged the student leader to offer more
solutions (M  6.00, SD  1.53), than participants in either the vision only condition (F(1, 81)  5.39,
p < .05, M  4.97, SD  1.85) or in the picture only condition (F(1, 81)  6.30, p < .05, M  4.85,
SD  1.81). No other effects were significant.
Attitudes
As intended and as in Experiment 2, the attitude towards the increase in study load, which is the focal
issue in this experiment, does not significantly correlate with either the attitude towards the increase in
involvement in university administration (r .13, ns), or with the attitude towards an increase in the
grant they receive (r .01, ns), while these last two do correlate with each other (r .27, p < .01).
We first analyzed the attitude towards the focal issue. The 2 (terror salience) Â 3 (charisma) ANOVA
on the focal attitude revealed an interaction effect, F(2, 81)  3.17, p < .05, h2
 0.07. Simple effects
analyses showed that charisma had no effect within the TV condition, F(2, 81)  0.19, ns, but it did
have an effect within the mortality salience condition, F(2, 81)  4.76, p < .05. Within the mortality
salience condition the picture  vision condition led to a more positive attitude towards an increase in
study load than either the vision only condition, F(1, 81)  6.97, p < .01, or the picture only condition,
F(1, 81)  7.01, p < .01 (See Figure 3).
In order to examine which conditions significantly differ from the baseline condition, we performed
a one-way ANOVA. This analysis with all seven conditions showed a trend for condition, F(6,
95)  1.95, p < .08, h2
 0.08. Only when participants were thinking about their own death and they
were presented with both a picture and vision (M  6.13, SD  1.25), there was a significant difference
(p < .05) from the baseline condition (M  4.13, SD  2.20), according to least significant differences
analyses.
We combined the two other attitude measures as they significantly correlate, r .27, p < .01. The 2
(terror salience) Â 3 (charisma) ANOVA on this combined attitude measure did not reveal any
significant effects (all F's < 2.15, all p's > .15). In order to examine which conditions significantly
differ from the baseline condition, we performed a one-way ANOVA. This analysis with all seven
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
400 Ernestine H. Gordijn and Diederik A. Stapel
conditions showed a significant effect for condition, F(6, 95)  2.34, p < .05, h2
 0.13. All conditions
(M  7.11; SD  1.11) significantly differ from the control condition (M  6.10; SD  1.11). This
suggests that providing a position in agreement with one's own opinion (within a further
counter-attitudinal message) strengthens one's opinion even more.
The findings from this third experiment show that when people experience an existential threat, in
this case after thinking about their own mortality, they are more likely to be influenced by a charismatic
leader with a mixed message, rather than when they are not reminded of such an existential threat. In
addition, the present findings indicate that neither attraction nor charismatic wording of the message
alone is sufficient to create charisma in our experiment. The leader was seen as most charismatic and as
offering more solutions when both characteristics were present. Moreover, only when the participants
were confronted with a leader who possessed both characteristics, did they change their attitude after
thinking about their own death.
EXPERIMENT 4
In the experiments thus far, we presented our participants with a controversial mixed message. In order
to test whether charismatic leaders are indeed only influential when they communicate a controversial
mixed rather than a one-sided message when people experience existential threat, we manipulated this
factor in the final study. Moreover, we aimed to show that mortality salience has similar effects as
salience of the possibility of terrorist attacks. Finally, we intended to show that not only attitudes are
influenced, but also behavior.
Method
Participants and Design
In this experiment 351 first-year psychology students at the University of Groningen participated
during a mass psychological testing session. The design was a 3 (Terror Salience: mortality salience vs.
terrorist attack salience vs. control: TV salience) Â 2 (Charisma: low vs. high) Â 3 (Message: mixed vs.
Figure 3. Attitude towards an increase in study load as a function terror salience and different aspects of charisma
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 401
one-sided vs. no message) factorial design. We also included a baseline condition (n  18) that received
no information.
Procedure
Participants were invited to the exam hall to fill out several questionnaires. The first questionnaire
concerned the manipulation of terror salience. Accordingly participants received either a mortality
salience manipulation (see Experiment 3), a terrorist attack salience manipulation (see Experiment 1),
or a TV salience manipulation (control: see Experiment 3) questionnaire. After this participants
received another relatively long questionnaire that was designed to manipulate charisma and message
characteristics; see Experiment 2.
The characteristics of the message were manipulated in the following way. In the mixed-message
condition, the student leader argued for an increase in study load (focal counter-attitudinal issue, i.e.,
the attitude that we intended to change), for more involvement of students in the university
administration (pro-attitudinal issue), and an increase in grants (pro-attitudinal issue). In the one-sided
message condition, the student leader pleaded besides for an increase in study load (focal
counter-attitudinal issue), also for less involvement of students in the university administration
(counter-attitudinal issue), and for a decrease in grants (counter-attitudinal issue). In the no message
condition, students only heard that the SBN wanted study load to be increased, but no further
information was given. In order to check for the effectiveness of our message manipulation,
participants were asked at the end of the questionnaire to what extent they thought that the message
pointed in one direction (1  absolutely not; 9  absolutely).
After reading this long paper, we measured the participants' attitudes. First we measured attitudes
towards the focal counter-attitudinal issue (i.e., `Study load should be increaseď), and the two other
issues (i.e., `Students should be more involved in university administration'; `Grants should be
increaseď). Participants responded on nine-point Likert scales (1  absolutely not; 9  absolutely).
These statements were embedded in a questionnaire that contained other irrelevant items about student
life.
After the manipulation check, we measured the participants' behavioral reaction towards the SBN.
Participants could respond in one of three ways: (1) by stating that they were not interested in the SBN,
(2) by stating that they wanted more information about the SBN, or (3) by stating that they wanted to
become a member. If they wanted more information, or wanted to become a member, they were also
asked to write their name and e-mail address on a separate piece of paper and to put it in an envelope
that we said we would send to the SBN for them (which of course we did not do, as the organization
does not exist). We also measured the participants' behavioral reaction towards the SBN in the control
condition. Because the baseline condition participants did not receive any information beforehand, they
were first told that there was a student organization called SBN. Afterwards participants were debriefed
and thanked.
Results and Discussion
Manipulation Check of Charisma
As expected, the 3 (terror salience) Â 2 (charisma) Â 3 (message) ANOVA on the measure of charisma
only revealed a main effect for charisma, F(1, 333)  10.03, p < .01, h2
 0.03. Results revealed that
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
402 Ernestine H. Gordijn and Diederik A. Stapel
participants in the charismatic condition judged the student leader to be more charismatic (M  5.22,
SD  1.54), than participants in the non-charismatic condition (M  4.64, SD  1.83).
Manipulation Check of Message
We measured to what extent participants believed that the message pointed in one direction. As
expected, the 3 (terror salience) Â 2 (charisma) Â 3 (message) ANOVA on this measure only revealed a
main effect for message, F(2, 333)  26.74, p < .01, h2
 0.14. Results revealed that participants in the
mixed message condition thought that the message pointed less in one direction (M  4.34, SD  1.33),
than participants in the one-sided condition (F(1, 333)  48.64, p < .01; M  5.72, SD  1.63), or in the
no-message condition (F(1, 333)  30.01, p < .01; M  5.43, SD  1.53)..
Attitudes
As intended, the attitude towards the increase in study load, which is the focal issue in this experiment,
did not correlate with either the attitude towards the increase in involvement in university
administration (r  .04, ns), or with the attitude towards an increase in the grant they receive (r .04,
ns), while these last two did correlate with each other (r .30, p < .01).
We first analyzed the attitude towards the focal issue. The 3 (terror salience) Â 2 (charisma) Â 3
(message) ANOVA on the focal attitude revealed a main effect for message, F(2, 333)  5.63, p < .01,
h2
 0.03, which suggests participants in the mixed message condition were more in favor of increasing
study load (M  3.74, SD  1.99), than participants in the one-sided condition (F(1, 333)  7.54,
p < .01; M  3.08, SD  1.77), or in the no-message condition (F(1, 333)  9.173, p < .01; M  2.92,
SD  1.81). The analysis further revealed a main effect for terror salience, F(2, 333)  3.05, p < .05,
h2
 0.02, which suggests participants in the TV condition were less in favor of increasing study load
(M  2.93, SD  1.99), than participants in the 9/11 condition (F(1, 333)  3.72, p  .06; M  3.40,
SD  1.77), or in the mortality salience condition (F(1, 333)  5.72, p < .05; M  3.50, SD  1.81). The
analysis finally revealed a main effect for charisma, F(2, 333)  19.78, p < .01, h2
 0.06, which
suggests participants in the charisma condition were more in favor of increasing study load (M  3.73,
SD  1.99), than participants in the non-charismatic condition (M  2.84, SD  1.71). No other effects,
including the expected three-way interaction, were significant, all F's < 1.
To test for our specific predictions regarding attitude change, we performed a one-way ANOVA in
order to examine what conditions significantly differ from the baseline condition in which only
attitudes and behavior were measured, which would suggest attitude change in the advocated direction.
This analysis with all 19 conditions showed a significant effect for condition, F(18, 350)  2.63,
p < .01, h2
 0.12. In line with our hypothesis, only the conditions in which a mixed message was
provided, the student leader was charismatic, and the participants were thinking either about 9/11
(M  4.55, SD  1.82), or about their own death (M  4.86, SD  1.62), showed a significant difference
(p's < .05) from the baseline condition (M  2.44, SD  1.20), according to least significant differences
analyses. All means are reported in Table 1.
We combined the two other attitude measures as they were significantly correlated. The 3 (terror
salience) Â 2 (charisma) Â 3 (message) ANOVA on this combined attitude measure only revealed a
main effect for message, F(2, 333)  32.41, p < .01, h2
 0.16. Results revealed that participants in the
mixed message condition agreed more with the increase in grants and involvement in university
administration (M  7.38, SD  1.29), than participants in the one-sided condition (F(1, 333)  49.46,
p < .01; M  6.24, SD  1.33), or in the no-message condition (F(1, 333)  47.91, p < .01; M  6.25,
SD  1.12).
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 403
In order to examine which conditions significantly differ from the baseline condition in which only
attitudes and behavior were measured we performed a one-way ANOVA. This analysis with all 19
conditions showed a significant effect for condition, F(18, 350)  4.32, p < .01, h2
 0.18. Only the six
conditions in which a mixed message was provided showed a significant difference (p's < .05) from the
baseline (M  6.28, SD  1.18), according to least significant differences analyses. This suggests that
after been given a message in which it is stated that both grants and involvement in university
administration should be increased leads to an even more positive attitude towards these issues than
participants already had, while arguing for the opposite, as in the one-sided condition, or not arguing for
it at all, as in the no message condition, does not lead to any change at all.
Behavioral Change
In order to measure to what extent the experimental conditions were successful in stimulating
participants to get more information or even become a member of the student organization (SBN),
participants were asked to fill out their names. We assigned to participants who did not sign up a 1, to
participants who wrote down their name on the leaflet to receive more information a 2, and to
participants who even wanted to become a member a 3. We performed a 3 (terror salience) Â 2
(charisma) Â 3 (message) ANOVA on this measure, which revealed a main effect for message, F(2,
333)  4.79, p < .01, h2
 0.03. This suggests that participants in the mixed message condition were
more likely to sign up (M  1.43, SD  0.55), than participants in the one-sided condition (M  1.26,
SD  0.44), or in the no-message condition (M  1.25, SD  0.44). The analysis further revealed a main
effect for charisma, F(1, 333)  3.58, p < .06, h2
 0.01, which suggests participants in the charisma
condition were more likely to sign up (M  1.37, SD  0.53), than participants in the non-charismatic
condition (M  1.26, SD  0.45). Results also revealed an interaction effect between charisma and
message, F(2, 333)  4.99, p < .01, h2
 0.03, and between charisma and terror salience, F(2,
333)  3.45, p < .05, h2
 0.02. But most importantly, there was the predicted three-way interaction
between charisma, terror salience, and message, F(4, 333)  2.36, p  .05, h2
 0.03. Simple effects
analyses revealed that the interaction between terror salience and message was not significant in the no
charisma condition, F(4, 333)  0.76, ns, while it was significant in the charisma condition, F(4,
333)  3.39, p < .01. Furthermore, within the charisma conditions, the effect of terror salience
significant in the mixed message condition, F(2, 333)  9.08, p < .01, but not in the one-sided
condition, F(2, 333)  0.06, ns, or in the TV condition, F(2, 333)  0.52, ns. Finally, within the
charismatic mixed message conditions, the death and 9/11 condition differ significantly from the TV
condition, F(1, 333)  14.65, p < .01 and F(1, 333)  13.12, p < .01, respectively. Means, which are
reported in Table 2, show support in line with our hypothesis: Only when participants are thinking about
terror-related events, such as their own death or 9/11, and who are confronted with a charismatic leader
Table 1. Focal attitude as a function terror salience, message characteristics, and charisma
9/11 Terror salience death Control
Message
One-sided Mixed Control One-sided Mixed Control One-sided Mixed Control
Charisma 3.6 (2.4) 4.6 (1.8) 3.5 (2.4) 3.5 (1.7) 4.9 (1.6) 3.5 (2.4) 3.2 (1.6) 3.4 (1.3) 3.3 (2.1)
No Charisma 2.9 (1.7) 3.3 (2.2) 2.6 (1.2) 2.9 (1.8) 3.5 (2.5) 2.8 (1.7) 2.6 (1.4) 2.8 (1.6) 2.3 (0.82)
Note: Standard deviations are reported between brackets. The mean of the baseline group is 2.4 (SD  1.2). The means in bold
differ significantly from the baseline group.
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
404 Ernestine H. Gordijn and Diederik A. Stapel
Table2.Behavioralreactionasafunctionofterrorsalience,messagecharacteristics,andcharisma
9/11TerrorsaliencedeathControl
Message
One-sidedMixedControlOne-sidedMixedControlOne-sidedMixedControl
Charisma1.2a(0.43)1.8b(0.64)1.3a(0.48)1.3a(0.45)1.8b(0.54)1.2a(0.38)1.2a(0.42)1.2a(0.55)1.3a(0.47)
NoCharisma1.1a(0.41)1.2a(0.31)1.1a(0.43)1.4a(0.50)1.3a(0.47)1.2a(0.41)1.3a(0.44)1.4a(0.59)1.2a(0.48)
Note:Meanswithdifferentsubscriptdiffersignificantlyfromeachotheratp<.01.Standarddeviationsarereportedbetweenbrackets.Themeanofthebaselinegroupis1.2
(SD0.43).Themeansinbolddiffersignificantlyfromthebaselinegroup.
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 405
who expresses a mixed message, are likely to sign up their name to get more information or even
become a member of the leader's organization.
In order to examine which conditions significantly differ from the baseline condition in which only
attitudes and behavior were measured we performed a one-way ANOVA. This analysis with all 19
conditions showed a significant effect for condition, F(18, 350)  2.74, p < .01, h2
 0.12. In line with
our hypothesis, only the conditions in which a mixed message was provided, the student leader was
charismatic, and the participants were thinking either about 9/11 (M  1.8, SD  0.64), or about their
own death (M  1.8, SD  0.54), showed a significant difference from the baseline (M  1.2,
SD  0.43), according to least significant differences analyses.
Taken together, the findings of this last experiment showed that charisma, salience of terror, and
mixed messages all contribute to greater persuasion. Although we did not find the predicted
three-way interaction with respect to the attitude, we did find, as expected, that a charismatic leader is
only effective in changing people's attitude in the advocated direction, when people are thinking
about terror-related events, such as their own mortality or the terrorist attacks, and when the leader
mixes his counter-attitudinal message with pro-attitudinal statements about another issue. People do
not only change their attitude in these circumstances, they are also more likely to show an interest in
the student leader's organization by requesting more information or even wanting to become a
member.
GENERAL DISCUSSION
As we explained in the introduction section of this paper, the starting point of our interest in the
psychology of charisma was formed by observations regarding the rise of Pim Fortuyn, a
controversial, charismatic political leader in The Netherlands. Initially, other politicians as well as
the general public did not take Pim Fortuyn's political ideas very seriously. However, much to the
amazement of many political commentators (Pels, 2003), Pim Fortuyn's political success increased
dramatically shortly after the terrorist attacks in the USA on 11 September 2001. Why? We analyzed
the Fortuyn case and reasoned that perhaps Fortuyn was so successful in post-9/11 times because he
was charismatic, and because the message he communicated included a smart mix of pro-attitudinal
and counter-attitudinal statements. That is, Fortuyn fulfilled people's need for direction and
solutions and, consequently, changed their attitudes, in a time of terror by presenting a vision to his
audience that mixed opinions with which they already agreed and things that were still controversial
to them.
These observations made us wonder whether we could generalize our analysis of this specific Pim
Fortuyn case to hypotheses concerning the conditions under which charismatic leaders in general are
most likely to be effective in changing the attitudes and behaviors of their audience. We hypothesized
that feelings of terror, caused by mortality salience or salience of terror-related events, provide a
charisma-conducive environment, which creates a need for vision, encouraging people to follow
charismatic leaders. Moreover, we hypothesized that when people experience such a high need for
vision due to feelings of terror, they are especially likely to be persuaded when the counter-attitudinal
message communicated by the charismatic leaders also includes some pro-attitudinal statements. Our
thought was that providing a counter-attitudinal message in combination with a pro-attitudinal
statement about another issue is a good strategy to avoid to be viewed as the out-group. In our opinion to
be seen as an in-group member, rather than as an out-group member, is important when facing
existential threat, because under those circumstances people search for others who share their cultural
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
406 Ernestine H. Gordijn and Diederik A. Stapel
worldviews (see e.g., Baldwin & Wesley, 1996). See Figure 1 which shows a graphic representation of
our model.
We tested our model in four experiments. As predicted, we found that existential threats such as fear
of terrorist attacks or mortality salience lead to an increased need for vision. Moreover, such an
increased need for vision opens people up to the influence of a controversial charismatic leader. This
leader, however, is only influential when he communicates a mixed rather than one-sided controversial
message.
Implications
These findings have implications for different areas of research in social psychology. First of all, we are,
to the best of our knowledge, the first to show in a controlled experiment that feelings of terror will
make a charismatic leader more influential (but see Cohen et al., 2004; Landau et al., 2004 for exploring
similar ideas; however, this research did not systematically study the influence of counter-attitudinal
messages and attitude change nor did it study why people follow charismatic leaders when
experiencing existential threat). Our findings support theorizing by Weber (1947) who argued that in
times of social crisis people are more open to charismatic leaders (see also, e.g., Becker, 1973; Bord,
1975; Friedland, 1964). According to, for example, Ehrhart and Klein (2001), such crises are
charisma-conducive environments that make followers more open to and receptive of the charismatic
qualities of the leader (see also, Bligh et al., 2004).
The question is what such charismatic qualities are? The (scientific) literature on charisma discusses
many factors that could influence the perception of charisma (see, e.g., Kirkpatrick & Locke, 1996). In
our research, we manipulated charisma by means of a charismatic wording of the message, as well as by
physical attractiveness. Our research suggests that to be perceived as charismatic and to be seen as
somebody who can offer solutions to the problems, it could help to be attractive. As Experiment 3
showed, the leader was most influential, and seen as most charismatic, when the leader was good
looking and spoke in a charismatic way.
Our findings also have interesting implications for Terror Management Theory (e.g., Greenberg
et al., 1986). This theory argues that when people are confronted with real terror (such as their own
mortality), they deal with this existential threat by searching for socially shared conceptions of reality
(a cultural worldview), and for a positive self-image. Dechesne and Kruglanski (2004) argued that in
this way people try to reduce the strong feelings of uncertainty they experience after being confronted
with this existential threat. Research has shown that this search for a cultural worldview and for a
positive self-image after experiencing existential threat has many consequences for the way people
feel, think, and act For example, Cohen et al. (2004) showed that mortality salience changes people's
preferences for task or relationship-oriented leaders to charismatic leaders. Furthermore, these
researchers showed that thoughts about terrorism increased people's preferences for George W. Bush,
who, they argue, has some charismatic qualities, rather than for his contender in the 2004 elections,
John Kerry. However, none of this research has investigated whether charismatic leaders can actually
change people's attitudes after experiencing existential threat. Moreover, previous research did not
investigate why charisma works in times of terror. Our research revealed that charismatic leaders are
more persuasive under circumstances that lead to existential threats, as such circumstances increase
people's need for vision, and vision is something provided by a charismatic leader.
Interestingly, our research indicates that fear of terrorism results in a general need for vision, and not
in a need for vision on how to deal with the specific problem (i.e., terrorism) that created their need for
vision. That is, after facing existential threats, such as one's own mortality or terrorist attacks, the
experiments showed that people followed a charismatic leader that offered a vision related to problems
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 407
concerning studying at the university. What would happen if the leader would have directly addressed
the terrorism threat? Such a leader might even be more persuasive as he literally, and not just
symbolically, alleviates people's existential concerns. However, the problem with providing direct
solutions to terrorism is that it will remind people of terrorism again, and thereby give rise to death
concerns. Future research could examine whether findings would even be stronger when the
charismatic leader would offer a vision on how to deal with the specific threat.
Another way in which our studies have implications for social psychological theorizing concerns
their relevance for persuasion research, in general. We showed that a persuasive counter-attitudinal
message is more influential when the message also includes some pro-attitudinal statements about
another issue. We argued that such a message could have more impact because people can agree with at
least part of the message, which therefore makes it easier for them to identify with the message. In this
way, the message can be attributed to the in-group, and thus the counter-attitudinal part will be seen as
more palatable, or at least worth considering (see e.g., Mackie et al., 1990, 1992). However, it is also
possible that the pro-attitudinal statements in mixed messages open people up to persuasion, because
they partly affirm people's values. This would be consistent with research by Sherman and Cohen
(2002) who showed that affirming some important values first leads to more persuasion. Alternatively,
it could be argued that presenting a mixed message makes the message itself more palatable or makes
the communicator appear more unbiased. Future research should further examine these different
explanations.
Interestingly, in both our studies we found that messages that include pro-attitudinal statements on
one issue (e.g., more involvement of students in the university administration) and counter-attitudinal
statements on another issue (e.g., enhancement of study load) lead to more favorable attitudes on both
these issues. That is, receiving a counter-attitudinal message that includes pro-attitudinal statements
about another issue (but no persuasive arguments supporting this other issue) made participants more
favorable to the counter-attitudinal part of the message as well as with the statements about the other
issue in the message that they already agreed with. On the other hand, people did not change their
attitudes when no pro-attitudinal statements regarding the other issue were included, or when these
statements about the other issue were also counter-attitudinal.
A reason for this `favorability' effect could be uncertainty reduction: knowing that another person
has similar opinions may strengthen your opinions because this means your opinions are not silly
(Festinger, 1954; see also Pyszczynski et al., 1996). It is also possible, however, that the favorability of
pro-attitudinal statements increases because of contrast: the presence of the counter-attitudinal
statements about the focal issue makes pro-attitudinal statements about the other issue look more
positive. As we did not include a condition that was pro-attitudinal with respect to both issues, we can
only speculate about which explanation is better. Future research should examine this question.
What is not clear from our research is whether people processed the information provided by the
charismatic leader in a systematic or heuristic manner (see, e.g., Chaiken, 1987). Some people might
argue that the fact that we obtained just main effects on the attitude measure in Experiment 1 indicates
that people used charismatic qualities such as communicator attractiveness and vision, and an unbiased
message as heuristic cues, which all increased agreement with the message. Indeed, all factors that we
manipulated can function as heuristic cues. However, if thinking about terror would merely lead to
openness to heuristic cues, we should have obtained only main effects of our manipulations on
behavior, which was not the case. The behavioral findings in Experiment 4 indicate that real change
only occurred when all factors were in place. Thus, only when a charismatic leader communicated a
mixed message to people who just thought about terror or their own death, these people were more
likely to join his organization. Still, what is not clear is whether people processed the leader's message
in a systematic or heuristic way. Future research may want to examine this by manipulating argument
quality.
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
408 Ernestine H. Gordijn and Diederik A. Stapel
CONCLUSION
The reason we started this research was that we wanted to further our understanding of a real life event,
namely the unexpected rise of Pim Fortuyn, a controversial, Dutch politician. We believe that our
analysis of the Pim Fortuyn case and the empirical, experimental studies on the psychology of
charismatic leadership in times of terror demonstrate that when social psychologists use their expertise
to address real life, the knife may cut both ways: our analysis strongly enhanced our understanding of a
real life and universally relevant phenomenon (people's reactions to political leaders and ideas in times
of terror) as well as advanced social psychological theory and research on charisma, terror management
theory, and persuasive communication.
Put differently, we would like to think that the current studies represent an approach to social
psychology that Blascovich (2003) feels is sorely needed when he lamented recently that `our field
seems to shy away from some important `big picture' issues; for example those pertinent to war, social
justice, politics, religion, etc.' (p. 9). Blascovich argues that social psychologists should promote the
practical value of the great theories that they develop. We agree. After all, `there is nothing so practical
as a good theory' (Lewin, 1951, p. 169). We believe that our social psychological analysis of the rise of
Pim Fortuyn adds to the understanding of why one man could have such an immense political influence
in such a short time. Moreover, the experimental studies that followed gave extra credence to our case
study analysis as well as added to our knowledge of the (social) psychology of charisma, management
of terror, and persuasion.
ACKNOWLEDGEMENTS
This research was supported by a `VIDI granť and a `Pionier Granť from the Dutch National Science
Foundation awarded to, respectively, the first and second author. Parts of this research have been
presented at the 2004 meeting of the Society of Personality and Social Psychology in Austin, Texas. We
thank David Marx, Mark Dechesne, and an anonymous reviewer for their very helpful comments on our
paper. We thank Marcel Zeelenberg for helping us with collecting the data of Experiment 2.
REFERENCES
Arndt, J., Greenberg, J., Solomon, S., Pyszczynski, T., & Schimel, J. (1999). Creativity and terror management.
The effects of creative activity and social projection following mortality salience. Journal of Personality and
Social Psychology, 77, 19­32.
Baars, M., Etman, E., Luitwieler, N., & Visser, M. (2002). Charismatisch politiek leiderschap; een vergelijkende
casestudy van Pim Fortuyn en Jan Peter Balkenende (Charismatic political leadership: A comparative casestudy
of Pim Fortuyn and Jan Peter Balkenende). Unpublished manuscript. Free University of Amsterdam.
Baldwin, M. W., & Wesley, R. (1996). Effects of existential anxiety and self-esteem on the perception of others.
Basic and Applied Social Psychology, 18, 75­95.
Bass, B. M. (1985). Leadership and performance beyond expectations. Mahwah, NJ: Lawrence E. Erlbaum
Associates.
Becker, E. (1973). The denial of death. New York: Free Press.
Belbin, R. M. (1981). Management teams: Why they succeed or fail. London: Heinemann (reprinted by Oxford:
Butterworth Heinemann).
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 409
Beyer, J. M. (1999). Taming and promoting charisma to change organizations. The Leadership Quarterly, 10,
307­330.
Blascovich, J. (2003). Rumors of the demise of personality and social psychology greatly exaggerated? Dialogue,
18, 9.
Bligh, M. C., Kohles, J. C., & Meindl, J. R. (2004). Charisma under crisis: Presidential leadership, rhetoric, and
media responses before and after the September 11th terrorist attacks. The Leadership Quarterly, 15, 211­239.
Bord, R. J. (1975). Toward a social-psychological theory of charismatic social influence processes. Social Forces,
53, 485­497.
Castano, E., Yzerbyt, V., Paladino, M. P., & Sacchi, S. (2002). I belong, therefore, I exist: Ingroup identification,
ingroup entitativity, and ingroup bias. Personality and Social Psychology Bulletin, 28, 135­143.
Chaiken, S. (1987). The heuristic model of persuasion. In M. P. Zanna, J. M. Olson, & C. P. Herman (Eds.), Social
influence: The ontario symposium (Vol. 5, pp. 3­40). Hillsdale, NJ: Erlbaum.
Cohen, F., Solomon, S., Maxfield, M., Pyszczynski, T., & Greenberg, J. (2004). Fatal attraction: The effects of
mortality salience on evaluations of charismatic, task-oriented, and relationship-oriented leaders. Psychological
Science, 15, 846­851.
Conger, J. A., & Karungo, R. A. (1987). Toward a behavioral theory of charismatic leadership in organizational
settings. Academy of Management Review, 12, 637­647.
Dechesne, M., Greenberg, J., Arndt, J., & Schimel, J. (2000). Terror management and the vicissitudes of sports fan
affiliation: The effects of mortality salience on optimism and fan identification. European Journal of Social
Psychology, 30, 813­835.
Dechesne, M., & Kruglanski, A. W. (2004). Terror's epistemic consequences: Existential threats and the quest for
certainty and closure. In J. Greenberg, S. Koole, & T. Pyszczynski (Eds.), Handbook of experimental existential
psychology. New York: Guilford Press.
Dumont, M., Yzerbyt, V., Wigboldus, D., & Gordijn, E. H. (2003). Social categorization and fear reactions to the
september 11th terrorist attacks. Personality and Social Psychology Bulletin, 29, 1509­1520.
Ehrhart, M. G., & Klein, K. J. (2001). Predicting followers' preferences for charismatic leadership: The influence
of follower values and personality. The Leadership Quarterly, 12, 153­179.
Ellemers, J. E. (2002). Pim Fortuyn: Een zuiver geval van charismatisch gezag (Pim Fortuyn: A clear case of
charismatic authority). Facta, 10, 2­5.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7, 117­140.
Florian, V., & Mikulincer, M. (1998). Terror management in childhood: Does death conceptualization moderate the
effects of mortality salience on acceptance of similar and different others? Personality and Social Psychology
Bulletin, 24, 1104­1112.
Friedland, W. H. (1964). For a sociological concept of charisma. Social Forces, 43, 18­26.
Gordijn, E. H., Postmes, T., & de Vries, N. K. (2001). Deviľs advocate or advocate of oneself: The effects of proand
counter-attitudinal advocacy and numerical support on self persuasion. Personality and Social Psychology
Bulletin, 27, 395­407.
Greenberg, J., Pyszczynski, T., Solomon, S., Rosenblatt, A., Veeder, M., Kirkland, S., & Lyon, D. (1990). Evidence
for terror management theory II: The effects of mortality salience on reactions to those who threaten or bolster
the cultural worldview. Journal of Personality and Social Psychology, 58, 308­318.
Greenberg, J., Solomon, S., & Pyszczynski, T. (1986). The causes and consequences of a need for self-esteem: A
terror management theory. In R. F. Baumeister (Ed.), Public self and private self. New York: Springer-Verlag.
Haslam, S. A., Platow, M. J., Turner, J. C., Reynolds, K. J., McGarty, C. M., Oakes, P. J., Johnson, S., Ryan, M. K.,
& Veenstra, K. (2001). Social identity and the romance of leadership: The importance of being seen to be `doing
it for us'. Group Processes and Interpersonal Relations, 4, 191­205.
House, R. J., Spangler, W. D., & Woycke, J. (1991). Personality and charisma in the U.S. Presidency: A
psychological theory of leader effectiveness. Administrative Science Quarterly, 36, 364­396.
Hunt, J. G., Boal, K. B., & Dodge, G. E. (1999). The effects of visionary and crisis-responsive charisma on
followers-A self-concept theory. The Leadership Quarterly, 10, 423­448.
Kirkpatrick, S. A., & Locke, E. A. (1996). Direct and indirect effects of three core charismatic leadership
components on performance and attitudes. Journal of Applied Psychology, 81, 36­51.
Klein, K. J., & House, R. J. (1995). On fire: Charismatic leadership and levels of analysis. Leadership Quarterly, 6,
183­198.
Landau, M. J., Solomon, S., Arndt, J., Greenberg, J., Pyszczynski, T., Miller, C. H., Cohen, F., & Ogilvie, R. M.
(2004). Deliver us from evil: The effects of mortality salience and reminders of 9/11 on support for president
George W. Bush. Personality and Social Psychology Bulletin, 30, 1136­1150.
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
410 Ernestine H. Gordijn and Diederik A. Stapel
Lewin, K. (1951). Field theory in social science; selected theoretical papers, D. Cartwright (Ed.). New York:
Harper & Row.
Mackie, D. M., Gastardo-Conoco, M. C., & Skelly, J. J. (1992). Knowledge of the advocated position and the
processing of in-group and out-group persuasive messages. Personality and Social Psychology Bulletin, 18,
145­151.
Mackie, D. M., Worth, L. T., & Asuncion, A. G. (1990). Processing of persuasive in-group messages. Journal of
Personality and Social Psychology, 58, 812­822.
Pels, D. (2003). De geest van Pim- Het gedachtegoed van een politieke dandy (The specter of Pim: The legacy of a
political dandy). Amsterdam, The Netherlands: AmbojAnthos uitgevers.
Pillai, R. (1996). Crisis and the emergence of charismatic leadership in groups: An experimental investigation.
Journal of Applied Psychology, 26, 543­562.
Pyszczynski, T., Greenberg, J., & Solomon, S. (1999). A dual-process model of defense against conscious and
unconscious death-related thoughts: An extension of terror management theory. Psychological Review, 106,
835­845.
Pyszczynski, T., Solomon, S., & Greenberg, J. (2003). In the wake of September 11: The psychology of terror.
Washington, DC: American Psychological Association.
Pyszczynski, T., Wicklund, R. A., Floresku, S., Koch, H., Gauch, G., Solomon, S., & Greenberg, J. (1996).
Whistling in the dark: Exaggerated consensus estimates in response to incidental reminders of mortality.
Psychological Science, 7, 332­336.
Rosenblatt, A., Greenberg, J., Solomon, S., Pyszczynski, T., Lyon, D. (1989). Evidence for terror management
theory: I. The effects of mortality salience on reactions to those who violate or uphold cultural values. Journal of
Personality and Social Psychology, 57, 681­690.
Shamir, B., & Howell, J. M. (1999). Organizational and contextual influences on the emergence and effectiveness
of charismatic leadership. The Leadership Quarterly, 10, 257­283.
Sherman, D. K., & Cohen, G. L. (2002). Accepting threatening information: Self-affirmation and the reduction of
defensive biases. Current Directions in Psychological Science, 11, 119­123.
Skitka, L. J., Bauman, C. W., & Mullen, B. (2004). Political tolerance and coming to psychological closure
following the September 11, 2001 terrorist attacks: An integrative approach. Personality and Social Psychology
Bulletin, 30, 743­756.
Weber, M. (1947). The theory of social and economic organization (A. M. Henderson, & T. Parsons (Eds and
Translator)). Glencoe, III: The Free Press.
Copyright # 2006 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 38, 389­411 (2008)
DOI: 10.1002/ejsp
Persuasive influence of charisma, terror, and mixed messages 411