595 Full

Group Processes & Intergroup
Relations
http://gpi.sagepub.com/
Leadership Efficacy and Women Leaders' Responses to Stereotype Activation

Crystal L. Hoyt and Jim Blascovich
Group Processes Intergroup Relations 2007 10: 595
DOI: 10.1177/1368430207084718
The online version of this article can be found at:

http://gpi.sagepub.com/content/10/4/595
Published by:
http://www.sagepublications.com
Additional services and information for Group Processes & Intergroup Relations can be found at:
Email Alerts: http://gpi.sagepub.com/cgi/alerts
Subscriptions: http://gpi.sagepub.com/subscriptions
Reprints: http://www.sagepub.com/journalsReprints.nav
Permissions: http://www.sagepub.com/journalsPermissions.nav
Citations: http://gpi.sagepub.com/content/10/4/595.refs.html
>> Version of Record - Nov 29, 2007
What is This?
Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Group Processes &
Intergroup Relations
2007 Vol 10(4) 595–616
Leadership Efficacy and

Women Leaders’ Responses
to Stereotype Activation
Crystal L. Hoyt
University of Richmond
Jim Blascovich
University of California, Santa Barbara
The role of leadership efficacy in women’s reactance responses to stereotype-based leadership

role expectations was examined in two laboratory studies. Participants, selected on the basis
of leadership efficacy scores, served as leaders of ostensible three-person groups. Half were
primed with the gender leadership stereotype. An immersive virtual environment designed for
this research served as the leadership setting. Results indicated that the effects of stereotype
activation on women leaders were moderated by leadership efficacy such that high efficacy
leaders exhibited more positive, reactance responses (increased perceived performance,
increased rated performance, greater domain identification, and higher well-being) than low
efficacy leaders. Additionally, perceived performance mediated the domain identification and
well-being effects of stereotype activation on high and low efficacy leaders.
keywords gender, leadership, reactance, self-efficacy, small groups, stereotypes
Organizational leaders make important Karau (2002) recently proposed an explanation

and far-reaching decisions that impact many for the dearth of women in top leadership
aspects of society. In the US relatively few of roles. According to their role congruity theory,
these powerful positions are held by women the descriptive and prescriptive aspects of the
who, despite increased representation in female gender stereotype are incongruent with
organizations, remain underrepresented in the leadership role, and that incongruity leads to
the upper echelons of American business and prejudice against women leaders. In addition to
government. For example, women represent leading to prejudice against women leaders, the
only 23.8% of Chief Executive Officers of all
organizations in the United States (Bureau of
Labor Statistics, 2006), and the numbers are Author’s note
even more bleak when we look to Fortune 500 Address correspondence to Crystal L.
companies where women hold a mere 7.9% of the Hoyt, Jepson School of Leadership Studies,
highest titles and there are only 11 women CEOs University of Richmond, Richmond, VA,
(2.2%; Catalyst, 2002; Joyce, 2006). Eagly and 23173, USA [email: choyt@richmond.edu]
Copyright © 2007 SAGE Publications

(Los Angeles, London, New Delhi and Singapore)
10:4; 595–616; DOI: 10.1177/1368430207084718
Group Processes & Intergroup Relations 10(4)
gender leader stereotype may also directly affect evidence indicates that stereotypically male
women in leadership positions. The goal of this qualities are thought necessary for successful
research is to understand how this stereotype leaders, and good leaders are described pre-
influences women in the leadership role. dominantly by masculine attributes (Arkkelin
& Simmons, 1985; Martell, Parker, Emrich, &
Crawford, 1998; Powell & Butterfield, 1979,
Gender stereotypes and the
1984, 1989; Rosenwasser & Dean, 1989; Schein,
leadership role 1973, 2001).
Stereotypic beliefs about gender differences
are pervasive, well documented, and highly Biases in the perception and evaluation of
resistant to change (Dodge, Gilroy, & Fenzel, female leaders
1995; Heilman, 2001). Gender stereotypes Individuals have incongruent expectations
encompass stereotypic beliefs about the attributes for female leaders: those based on gender
of women and men that prescribe how men stereotypes and those based on leadership
and women should, or ought to, be1 (Burgess & roles. The perceived incongruity between the
Borgida, 1999; Glick & Fiske, 1999). Descriptive female gender role and the leadership role
stereotypical attributes of men include ‘agentic’ leads to prejudice that can account for the
characteristics emphasizing confidence, control, numerous findings that indicate less favorable
and assertiveness whereas stereotypical attributes attitudes toward female than male leaders,
of women include ‘communal’ characteristics greater difficulty for women to attain top
highlighting a concern for others (Broverman, leadership roles, and greater difficulty for these
Vogel, Broverman, Clarkson, & Rosenkrantz, women to be viewed as effective in these roles
1972; Eagly, 1987; Eagly, Wood, & Diekman, (Eagly & Karau, 2002). For example, Heilman
2000; Heilman, 2001). Although women and men and colleagues have shown that while women
are viewed as differing on other traits (Deaux managers are seen as more similar to successful
& Lewis, 1983, 1984; Eckes, 1994), the agentic managers than women in general, they are rated
and communal stereotypical attributes directly as more different from successful managers
relate to the leadership domain. Substantial em- than are male managers (Heilman, Block, &
pirical research supports the prescriptive nature Martell, 1995; Heilman, Block, Martell, & Simon,
of gender stereotypes (Glick & Fiske, 1996; 1989). Researchers have also demonstrated that
Williams & Best, 1990; Wood, Christensen, women are presumed to be less competent leaders
Hebl, & Rothgerber, 1997). Accordingly, not than men and less worthy of the leadership posi-
only should individuals conform to their stereo- tion across a variety of contexts (Boldry, Wood,
typic description, but they should also avoid be- & Kashy, 2001; Carli & Eagly, 2001; Heilman,
haviors that are incompatible with the stereotype 2001; Ridgeway, 2001; Schein, 2001). Addi-
(Heilman, 2001). tionally, research has revealed that in order to
Role congruity theory maintains that the agen- be influential leaders, women need to combine
tic qualities deemed necessary in the leadership communal qualities (e.g. warmth and friendliness)
role are incompatible with the predominantly with agentic qualities (e.g. competence and
communal qualities associated with women directiveness; Carli, 2001; Eagly, Makhijani, &
(Eagly & Karau, 2002). This gender leader Klonsky, 1992; Rudman & Glick, 2001). While
stereotype (aka think-manager-think-male empirical research clearly establishes the impact
stereotype) is well validated (Sczesny, 2003). of stereotypes on the perception and evaluation
Top management positions and executive level of women leaders, the literature is less clear on
jobs are almost always thought to require an the impact of these stereotypes on women leaders
achievement-oriented aggressiveness and an themselves; the goal of the current research is
emotional toughness that is antithetical to the to examine this impact.
female gender stereotype. Substantial empirical
596

Hoyt & Blascovich leadership efficacy, stereotype activation, and women leaders
The impact of stereotype activation to possess sufficient power in order to react

on the target against the stereotype (Kray, Reb, Galinsky, &
Thompson, 2004).
Responses to stereotypes vary from deleterious This distinction between stereotype vulner-
threat responses to more benign reactance ability and stereotype reactance is similar to the
responses. Initial research into the impact of distinction between stereotype assimilation and
stereotypes on targets generally demonstrated contrast (Stoddard, Kliengklom, & Ben-Zeev,
the pernicious effects of stereotype threat, 2003). Assimilation effects refer to performance
defined as the apprehension that individuals feel deficits resulting from stereotype threat. Contrast
when they are at risk of confirming a negative effects refer to increases in performance by indi-
stereotype about their group (Aronson, Quinn, viduals primed with negative stereotypes re-
& Spencer, 1998). Stereotype threat undermines garding their group. Stoddard and colleagues
individuals’ assessments of their abilities and examined the moderating role of the subtlety of
results in decreased performance; it has been the stereotype prime on women’s desire to take
shown to play an important role in the under- on a leadership role. Specifically, they found that
performance of minorities (Aronson et al., 1998; subtle stereotype activation evoked assimilation
Steele & Aronson, 1995). Stereotype vulner- such that women were less likely to desire a
ability effects have been demonstrated in the leadership position, whereas blatant stereotype
leadership domain. Davies, Spencer, and Steele activation resulted in a heightened desire to
(2005) found that exposure to stereotypic assume a leadership position.
commercials undermines women’s leadership
aspirations; however, presenting women with
The role of efficacy in the impact of
an identity safe environment eliminates this
vulnerability. Additionally, Bergeron, Block, & stereotype activation
Echtenkamp (2006) found that men outper- Self-efficacy, a key construct derived from
formed women on a managerial task but only Bandura’s social-cognitive theory (1986), is
when it was a masculine sex role-typed task; defined as ‘belief in one’s capabilities to organize
this effect disappeared on a feminine sex role- and execute the courses of action required to
typed task. produce given attainments’ (Bandura, 1997,
As is clear from the responses of women in p. 3). Empirical studies yield consistent findings
identity safe environments (Davies et al., 2005) revealing that self-efficacy influences thought
and feminine sex-typed tasks (Bergeron et al., patterns, stress reactions, and coping; what
2006), stereotypes are not always met with people choose to do; their persistence in the face
vulnerability responses. Indeed, research has of difficulty; how much effort they put forth; and
shown that when gender stereotypes are primed performance (Bandura, 1982, 1986; Bandura &
explicitly women sometimes respond by engag- Cervone, 1983; Bandura & Wood, 1989; Barling
ing in stereotype countering behaviors; that is, & Beattie, 1983; Campbell & Hackett, 1986; Hill,
they demonstrate stereotype reactance (Kray, Smith, & Mann, 1987; Lent, Brown, & Larkin,
Thompson, & Galinsky, 2001). Brehm’s (1966) 1987; Taylor, Locke, Lee, & Gist, 1984).
psychological reactance theory contends that
people respond to perceived threats to their Self-efficacy as a moderator of stereotype
freedom (such as stereotyped expectations of activation effects
inferiority) by attempting to reassert their Self-efficacy has been shown to moderate
freedom (engaging in counterstereotypical be- responses to work-related stressors and is
havior). For example, Kray and colleagues (2001) associated with the ways in which individuals
showed that women blatantly presented with the cope with stressors (Jex & Bliese, 1999; Stumpf,
gender and bargaining stereotype outperform Brief, & Hartman, 1987). Individuals high in
men at the bargaining table. Additionally, Kray self-efficacy adopt a more problem-focused
and colleagues found that participants need coping strategy; those lower in self-efficacy
597

assume a more emotion-focused approach. perseverance, resilience in the face of adversity,

These approaches are not equally effi cient; higher accomplishments, and success at new
problem-focused coping is associated with activities (Bandura, 1997). Consequently, we
more adaptive responses than emotion-focused hypothesized that leadership efficacy would
coping (Lazarus & Folkman, 1984). Construing interact with stereotype activation such that
stereotype activation as a potential stressor, high efficacy leaders would report higher levels
reactions to stereotype activation seem likely to of perceived performance when the stereotype
be moderated by levels of self-efficacy. was primed than when neutral material was
Recent research supports the prediction that primed. Additionally, based on the stereotype
self-efficacy moderates responses to stereotype reactance literature we predicted that ratings
activation. Solo status, or being the lone woman of performance would show that stereotype
in a group of men, is a sufficient situational activation increases the performance of high
cue to induce stereotype activation for women efficacy leaders compared to the control
(Inzlicht & Ben-Zeev, 2000; Sekaquaptewa & condition.
Thompson, 2003). White and Gardner (2003)
examined the impact of task-specific efficacy on Domain identification
the effect of solo status on math performance. Stereotype threat theory proposes that one
They found that women’s efficacy moderated way individuals may cope with the threat is
their performance on a math test when they by disidentifying with the stereotype-relevant
were the solo female group member. Specifically, domain. That is, individuals’ self-perceptions and
highly confident women performed better in the self-concept become independent of perform-
stereotype activating solo condition than less ance in that domain. Such disidentification can
confident women. In the current research, self- diminish achievement in the domain (Steele &
efficacy for leadership is predicted to moderate Aronson, 1995). For example, stigmatized indi-
women’s responses to stereotype activation in viduals (e.g. African–Americans) are more prone
the leadership domain such that those with high than their non-stigmatized counterparts to dis-
levels of efficacy will demonstrate reactance identify with academics, academic performance
responses. situations being a domain in which stereotype
threat often occurs (Major, Spencer, Schmader,
Wolfe, & Crocker, 1998). Consequently, we
Hypotheses
hypothesized that leadership efficacy would
In the current research, stereotypes were interact with stereotype activation such that
primed in a blatant manner, and self-efficacy high efficacy leaders would show reactance by
was examined as a potential moderator of reporting more identification with leadership
stereotype reactance effects.2 We hypothesized in the stereotype activation compared to the
that leadership effi cacy would interact with control condition.
stereotype activation on the following responses:
perceived and rated performance, domain Psychological well-being
identification, and psychological well-being. Previous research has shown that negative
stereotypes can adversely affect the psychological
Perceived and rated performance well-being of the targets (Swim, Hyers, Cohen,
In situations, such as leadership situations, that & Ferguson, 2001). Psychological well-being
often lack clear and instantaneous measures of includes, but is not limited to, reported levels of
effectiveness, individuals’ perceptions of how self-esteem and depressed affect. Consequently,
well they performed become a very important we hypothesized that leadership efficacy would
source of immediate feedback to the self. interact with stereotype activation such that high
Perceptions of successful performance can efficacy leaders would report greater levels of
convince people that they have what it takes to well-being in the stereotype activation compared
succeed in a domain which can in turn result in to the control condition.
598

Mediational hypotheses a paradigm designed to test the hypotheses in

Based on self-perception theory and self-esteem a controlled, small group, leadership situation.
theory, self-perception was hypothesized to medi- Given that the primary goal of these studies
ate the proposed interaction effects on domain was to demonstrate the differential effects of
identifi cation and psychological well-being. stereotype activation on high and low efficacy
First, according to Bem’s self-perception theory leaders, participant selection was based on
(1967), individuals infer their own traits and leadership efficacy. Participants took part in a
attitudes by observing their own behavior. Thus, group task as the ‘randomly assigned’ leader of
we hypothesized that individuals’ perceptions a three-person group. Half of the participants
of their leadership performance would influence were primed with the gender leader stereotype
their inference about how strongly they view before taking the leadership role in an employee
themselves as leadership-oriented persons; that hiring decision task.
is, how much they identify with the domain of
leadership. Second, we hypothesized percep- Immersive virtual environment technology
tions of performance on the stereotype domain Historically, experimental investigations into
relevant task would directly influence individuals’ leadership and other facets of group life have
well-being. That is, perceiving that one performed employed impactful manipulations to induce
well on the leadership task and thus disconfirmed participants to experience certain affects, cog-
the negative stereotype would lead to increased nitions, and/or motivations. Common simu-
psychological well-being (see Figure 1 for lation techniques include the use of vignettes,
hypothesized mediational relationships.) role-playing, confederates, and verbal instruc-
tions. Recent technological advances have
provided researchers with a tool to facilitate
Overview of present research compelling manipulations: immersive virtual
Procedural overview environment technology (IVET). We define a
These hypotheses were addressed in two experi- virtual environment as one providing synthetic
mental studies. Each of these studies employed sensory information that leads to perceptions of
Figure 1. Perceived performance as a hypothesized mediator of the interactive effects of stereotype activation
and leadership efficacy on domain identification, self-esteem, and depressed affect.
599

environments and their contents as if they are

not synthetic (Blascovich et al., 2002).
Although relatively new, IVET technology has
been noted for many of the methodological
advantages it provides (see Blascovich et al.,
2002 for a detailed description). First, it pro-
vides superb control over the confederates’
(followers’) demographic characteristics, such as
sex and ethnicity, and behaviors, assuring exact
replications of confederates’ behaviors across
participants. Additionally, IVET diminishes
the need for live research confederates, thus
increasing efficiency in conducting experi-
ments. In terms of operation, our IVET system
consisted of three subsystems: a) a body location
and head tracker, b) a graphics rendering
computer, and c) an audiovisual head-mounted
display (HMD). Translation and orientation
information regarding the user’s body and head
is sensed and recorded by the trackers, which
in turn inform the rendering computer, which
then generates and projects visual and auditory
stimuli to the user’s HMD3 (see Figure 2).
These experiments were conducted in a lab-
oratory that contained three rooms in which
three people, each located in a separate room,
can interact together within a virtual conference
room consisting of three people sitting around
a table. The leadership task took place within
this immersive virtual conference room with,
what the leaders were led to believe were, two
other research participants, Chris (male) and
Michelle (female). Participants were led to
believe that the two other participants would don
their HMDs in the two other laboratory rooms
and that they would all be networked together
into one virtual conference. In the virtual room
Figure 2. The leader’s view of the two followers
the participant/leader sat across the conference during the meeting in the immersive virtual
table from the two ostensible group members environment and the immersive virtual display
who were programmed to generate random non- system.
verbal body movements, including eye blinks,
head movements, and small facial expressions leadership per formance, and leadership
(see Figure 2). domain identification and that the latter
moderation effect would be mediated by
perceived performance.
Study 1
This study tested the hypotheses that leadership Method
efficacy would interact with stereotype activation Participants and design Fifty-four female
on perceived leadership performance, rated undergraduate students enrolled in Introductory
600

Psychology at a major research university served three-minute meeting with the other two ‘par-
as participants and were given either credit ticipants’. The participants were then immersed
toward the Introductory Psychology course in a virtual world for a three-minute meeting, sup-
research participation requirement or US $10 posedly with the other participants.4 Following
for their participation. Of the 54, one refused this, the participants completed the perceived
to perform the leadership task, leaving 53 par- performance and domain identification ques-
ticipants in the final sample. The experiment tionnaires, while the other group members were
employed a 2 (Leadership Efficacy: High or ostensibly completing the task. Participants were
Low) × 2 (Stereotype Activation: Primed or Not) then thoroughly debriefed, thanked, and given
between-subjects quasi-experimental design. course credit or payment.
Participant selection: Self-efficacy for leadership Stereotype activation manipulation To prime

(SEL) Participants were selected on the basis the gender leader stereotype, participants were
of SEL scores obtained during a prescreening given a folder to peruse while waiting for the
session at the beginning of the term. Murphy experimenter to return. For the stereotype activ-
(1992) developed the SEL to measure individuals’ ation condition, the folder contained media
self-efficacy regarding their general leadership images of male leaders and information re-
abilities. The SEL has been well-validated in garding the gender gap in top leadership roles;
predicting leadership, group, and organizational for example, participants were told that ‘in nearly
outcomes (Chemers, Watson, & May, 2000; all major corporations, the military and political
Murphy, 2002; Murphy, Chemers, Kohles, & offices, men outnumber women in top leadership
Macaulay, 2004; Watson, Chemers, & Preiser, roles, and there seems to be a steady gap in men
1996). Participants rated their leadership abil- and women in top leadership roles. Though women
ities on eight 5-point Likert-type scales from 1 represent 46% of the U.S. workforce, they hold only
(strongly disagree) to 5 (strongly agree). Example about 6% of executive titles like CEO, chairwoman,
statements include ‘I am confident of my abil- and executive vice president. And only four women
ity to influence a work group that I lead,’ and hold CEO positions at Fortune 500 companies’.5 In
‘I know what it takes to keep a work group run- the control condition, the folder contained
ning smoothly’ (α = .93). Those scoring in the images of and information regarding the virtual
upper and lower quartiles of the distribution of reality lab. To augment the stereotype activation
prescreening scores were eligible for participation manipulation, participants in the experimental
in this study. condition were also told by the experimenter that
the research was aimed at better understanding
Experimental over view Participants were gender differences in leadership abilities; those
randomly assigned to receive the stereotype or in the control condition were simply told that
the neutral prime. All participants ostensibly the research was designed to better understand
participated in a group task as the randomly leadership abilities.
assigned leader of the three-person group. Par-
ticipants were run individually. Upon arrival, Leadership task scenario Participants occupied
participants were greeted by one of two male ex- the role of President of the Human Resources
perimenters. They were told that they were ran- division of James Frick Inc., whose charge was
domly selected to be the group leader and were to lead (i.e. chair) a selection committee hiring
asked to arrive earlier than the other two group a new junior associate. Specifically, they were
members. After this general introduction to the assisted by two Vice-Chairs; that is, the other
experiment, the stereotype activation mani- two ‘participants’. Leaders were given a memo
pulation (see below) was introduced. The from the CEO of James Frick Inc. explaining
participants (i.e. leaders) then received instruc- the need for the new associate in detail and
tions regarding the group task and were given were given two applicant information packets.
seven minutes to prepare for an audio-recorded They were informed that the two Vice-Chairs
601

would be given the same memo and packets of measures, the following items were created to
applicant information after the meeting. The assess leadership identification: ‘Leadership
participants’ responsibility was to examine the is important to me’, ‘I am a leadership-oriented
materials and prepare for a three-minute meeting person’, ‘It is important for me to be selected group
with the Vice-Chairs. Their job was to explain leader’, ‘I am a good leader’, and ‘Leadership skills
the task to the followers and advise them on will be important to my career’ (α = .88).
how they thought the followers should go about
accomplishing the task. More generally, they Manipulation check To assess the efficacy of the
were asked to influence and motivate them to stereotype activation manipulation, participants
make the best hiring decision possible. Within were asked to rate the following statement: ‘In
an immersive virtual conference room, the leadership roles, people f my gender often face
leaders held a three-minute meeting with the biased evaluations’.
two Vice-Chairs; this meeting allowed for only
one-way communication from the leader to
Study 2
the followers. This task was fashioned after the
increasingly common virtual workplace. Due, in part, to the small sample size for Study 1
(n = 53), we conducted Study 2 as a replication
Measures Participants responded to all meas- as well as an extension of Study 1. The methods
ures on a 7-point Likert-type scale ranging from for both studies are presented and, for simplifi-
–3 (strongly disagree) to 3 (strongly agree). cation, the data from both studies are analyzed
together. Study 2 was extended to test well-being
Perceived performance Leaders were asked to rate responses. Thus, efficacy was expected to interact
their own performance on two items. The with stereotype activation on perceptions and
following items were created to assess the par- ratings of leadership performance, identifica-
ticipants’ perceived performance: ‘I performed tion with the domain of leadership, and well-
well on the leadership task I just completed’, being responses. Additionally, this study tested
and ‘I am confident that I performed well’ the prediction that the interaction between
(α = .87). leadership efficacy and stereotype activation
on domain identification and well-being would
Rated performance All audiotapes were inde- be mediated by perceived performance.
pendently coded by two trained raters blind to
leadership efficacy and stereotype activation Method
condition. The rated performance scale consisted Participants and design Seventy-five female
of five items: vigor (how energetic, active, and undergraduate students attending a major
lively they were), anxiety (nervousness; reverse research university participated in this study.
coded), task explanation (how well they explained Participants were given either credit toward
the problem), vision (how well they articulated the Introductory Psychology course research
a clear vision for the company), and authority participation requirement or US $10 for their
(how knowledgeable, direct, and down to busi- participation. Of the original 75 participants,
ness they were). Performance assessments were 3 were dropped from analyses for either not
made on a 9-point scale. Interrater reliability understanding the experimental instructions
was computed using a Pearson’s r (r = .89, or not fluently speaking English. Thus, there
p < .001). were 72 participants in the final sample. The
design of Study 2 was identical to that of Study
Domain identification Domain identification 1 and participants were selected in the same
was operationalized as the extent to which manner.
the participants identified with being, and saw
themselves as, a leader. Fashioned after Steele Procedures The procedures for Study 2 were
and Aronson’s (1995) academic identification basically the same as in Study 1 with two minor
602

changes. First, both experimenters in Study 1 were pessimistic about the future’, ‘I feel sad’, ‘I feel
male; four female and one male experimenter mortified’, and ‘I feel hopeless’ (α = .85).
were employed in the current study. Additionally,
the experimental instruction set was standardized Results
such that most instructions were given to the The manipulation check and the first set of
participants via an audio-recording rather than moderation hypotheses were tested with the
via the experimenter. The remainder of the data from both studies. To test the hypotheses
methodology remained identical to Study 1; dependent variables were analyzed with a series
the stereotype activation manipulation, the of 2 (Leadership Efficacy) × 2 (Stereotype
immersive virtual environment, and the task Activation) × 2 (Study) between-subjects an-
were the same across studies. alyses of variance. Study was included as a
factor in the analyses to ensure equivalence
Measures Again, participants responded to in the data from the two studies.
all measures on a 7-point Likert-type scale rang-
ing from –3 (strongly disagree) to 3 (strongly Manipulation check As expected, participants
agree). in both studies who received the stereotype
activation manipulation were more likely to
Perceived performance, rated performance, domain report that people of their gender face biased
identification, and manipulation check The per- evaluations (M = 1.26, SD = 1.20) than those
ceived performance and domain identification who did not receive the stereotype manipulation
measures remained identical to those in Study (M = 0.57; SD = 1.58) (F(1, 117) = 8.04, p < .01,
1 (α = .93 and α = .90, respectively). Also, the η2 = .06). The factor of study did not have a
performance rating procedures were identical main effect nor did it interact with condition
(r = .80, p < .001) and the same manipulation or leadership efficacy on the manipulation
check item was used. check. There was no main effect for leadership
efficacy (p > .35) and there was no interaction
Self-esteem Self-esteem was assessed with a between leadership efficacy and stereotype
6-item measure adapted from Heatherton activation (p > .45).
and Polivy’s (1991) state self-esteem measure
(α = .85). Three items assessed performance Moderation hypotheses: Perceived
self-esteem: ‘I feel frustrated or rattled by my performance, rated performance,
performance’, ‘I feel that I am having trouble and domain identification
understanding things’, and ‘I feel like I am Perceived performance The Cronbach’s alpha of
not doing well’. The remaining three items the perceived performance scale for the com-
assessed social self-esteem: ‘I feel displeased bined data was .91. These analyses revealed that
with myself’, ‘I am worried about looking the factor of study did not have a main effect
foolish’, and ‘I feel concerned about the im- nor did it interact with the two independent
pression I am making’. Both performance and variables on perceived performance. Addition-
social self-esteem resulted in similar outcomes, ally, there was no main effect for stereotype
so, for simplicity, we have combined the measures activation on perceived performance (p > .60).
into a general self-esteem measure. There was a main effect of efficacy such that
high efficacy women perceived they performed
Depressed affect The depressed affect scale con- better (M = 0.78, SD = 1.42) than low efficacy
sisted of six items adapted from the Multiple women (M = –.64, SD = 1.41) (F(1, 117) = 28.52,
Affect Adjective Check List (MAACL) (Lubin, p < .001, η 2 = .20). Additionally, there was a
Zuckerman, & Woodward, 1985). Participants significant interaction (F(1, 117) = 9.08, p < .01,
indicated their agreement to the following items: η2 = .07; see Figure 3) such that high efficacy
‘I feel discouraged’, ‘I feel distressed’, ‘I feel women primed with the stereotype reported
603

higher levels of perceived performance (M = 1.17,

SE = 0.25) compared to those primed with the
neutral material (M = 0.29, SE = 0.27) (simple
F(1, 117) = 5.74, p < .05, η2 = .05). Additionally,
low efficacy women primed with the stereotype
reported marginally lower levels of perceived
per formance (M = –0.94, SE = 0.24) than
those primed with the neutral material (M =
–0.30, SE = 0.25; simple F(1, 117) = 3.42, p < .07,
η2 = .03).
Figure 3. The impact of leadership efficacy and

stereotype activation on perceived leadership
performance, rated leadership performance, and
domain identification.
Rated performance Unfortunately, a few of the

tapes were inaudible, leaving 109 data points
for these analyses. Interrater reliability for the
combined data was r = .83, p < .001. The factor
of study did not have a main effect nor did it
interact with the two independent variables on
rated performance. Additionally, there was no
main effect for stereotype activation on rated
performance (p > .15). There was a main effect
of efficacy such that high efficacy women were
rated as performing better (M = 5.51, SD = 1.63)
than low efficacy women (M = 4.48, SD = 1.85)
(F(1, 101) = 9.68, p < .05, η2 = .09). Additionally,
there was a marginally significant interaction
(F(1, 101) = 2.90, p = .09, η2 = .03; see Figure 3)
such that high efficacy women primed with the
stereotype had higher levels of rated performance
(M = 6.05, SE = 0.35) compared to those primed
with the neutral material (M = 5.02, SE = 0.35)
(simple F(1, 101) = 4.28, p < .05, η2 = .04).
Domain identification The Cronbach’s alpha of
the domain identification scale for the combined
data was .89. The study factor did not have a
main effect on domain identification, but it did
interact with efficacy on domain identification
(F(1, 117) = 7.94, p < .01). The interaction was
Figure 3 Continued such that high efficacy leaders had higher domain
604

identification in Study 2 than Study 1, and low

efficacy leaders had lower identification in Study
2 than Study 1; in other words, the main effect of
efficacy on domain identification was stronger in
Study 2. The main effect of stereotype activation
on domain identification was not significant
(p > .50). There was a significant main effect of
efficacy (F(1, 117) = 105.32, p < .001, η2 = .47),
with high efficacy women identifying more with
the domain of leadership (M = 1.72, SD = 0.88)
than low efficacy women (M = –.15, SD = 1.07).
Also, there was a significant interaction between
leadership efficacy and stereotype activation on
domain identification (F(1, 117) = 5.91, p < .05,
η2 = .05) (see Figure 3); high efficacy women
primed with the stereotype reported higher levels
of domain identification (M = 1.93, SE = 0.17)
than those primed with the neutral material
(M = 1.41, SE = 0.18; simple F(1, 117) = 4.28,
p < .05, η2 = .04). Low efficacy leaders primed
with the stereotype reported similar levels of
domain identification (M = –.27, SE = 0.16)
to those in the control condition (M = 0.05,
SE = 0.17) (simple F(1, 117) = 1.82, p = .18).
Moderation hypotheses: Self-esteem and

depressed affect Data to test these hypotheses
were only collected in Study 2. To test the
hypotheses we analyzed the dependent variables
with a series of 2 (Leadership Efficacy) × 2
(Stereotype Activation) between-subjects analyses
of variance.
Self-esteem The main effect of stereotype activ- Figure 4. The impact of leadership efficacy and stereotype
ation on self-esteem was not significant (p > .15). activation on self-esteem and depressed affect (Study 2).
There was a significant main effect of efficacy
(F(1, 67) = 16.12, p < .001, η2 = .19), with high stereotype did not report significantly different
efficacy women reporting higher self-esteem levels of self-esteem (M = –0.47, SE = 0.28) com-
(M = 0.79, SD = 1.35) than low efficacy women pared to those in the control condition (M = –0.25,
(M = –0.36, SD = 1.11). There was a significant SE = 0.28), simple F(1, 67) = .32, ns).
interaction between leadership efficacy and
stereotype activation (F(1, 67) = 5.10, p < .03, Depressed affect The main effect for stereotype
η2 = .07; see Figure 4) such that high efficacy activation on depressed affect was not significant
women primed with the stereotype reported (p > .80). There was a significant main effect for
higher levels of self-esteem (M = 1.30, SE = 0.28) efficacy (F(1, 66) = 17.04, p < .001, η2 = .21), such
than those primed with the neutral material that high efficacy women reported less depressed
(M = 0.25, SE = 0.29; simple F(1, 67) = 6.85, p < .02, affect (M = –2.52, SD = 1.17) than low efficacy
η2 = .09). Low efficacy leaders primed with the women (M = –1.11, SD = 1.69). As predicted,
605

there was a significant interaction between and finally, the interaction term was entered in
leadership efficacy and stereotype activation the third step. The next criterion was to show that
on the leaders’ self-reported depressed affect the mediator variable was significantly related
(F(1, 66) = 4.48, p < .04, η2 = .06 ) (see Figure 4). to the dependent variable. Consistent with this
Testing the simple effects of stereotype activ- criterion, in Step 2 perceived performance
ation revealed no significant effect of stereotype significantly predicted domain identification
activation on self-reported depressed affect (β = .32, p < .001). The final criterion for medi-
for either high efficacy (simple F(1, 66) = 1.80, ation is that the original interaction term should
p > .15) or low efficacy (simple F(1, 66) = 2.74, be substantially reduced or eliminated when the
p > .10) women. However, testing the simple mediator is entered into the analysis. Indeed,
effects of leadership effi cacy indicated that in Step 3 the interaction term was no longer
although in the control condition high and significant ( β = .08, p > .15) and perceived
low efficacy women report similar levels of performance significantly predicted domain
depressed affect (simple F(1, 66) = 2.02, p > .15), identification (β = .30 p = .001; refer to Table 1
when primed with the stereotype high efficacy and see Figure 3).6
women reported significantly lower levels of Additionally, using the raw regression coeffi-
depressed affect (M = –2.85, SD = .35) than low cients and the standard errors from the analyses
efficacy leaders (M = –0.71, SD = 0.34) (simple above, the Goodman (I) version of the Sobel
F(1, 66) = 19.49, p < .001, η2 = .23). test statistic was calculated to test whether the
effect of the interaction between leadership
Mediation hypotheses efficacy and stereotype activation on domain
Domain identification Data from both studies identification via perceived performance is
were used in this analysis. Regression techniques significantly different from zero (Baron &
were used to test whether the hypothesized Kenney, 1986; Sobel, 1982.) The Sobel test
significant interaction between stereotype statistic was significant (Z = 2.37, p < .02), indi-
activation and leadership efficacy on domain cating that perceptions of performance seem
identification was mediated by perceived per- to carry the interactive influence of leader
formance (Baron & Kenny, 1986). To test this efficacy and stereotype activation on domain
mediated moderation, the interaction term was identification.
computed and domain identification was re-
gressed on the main effects and the interaction Self-esteem Next, the hypothesis that the inter-
of stereotype activation and leadership efficacy action between stereotype activation and leader-
to determine the total percentage of variance ship efficacy on self-esteem would be mediated
explained by the interaction. As expected from by perceived performance was tested (refer to
the analysis of variance results above, the Table 1); self-esteem data were collected only in
interaction coefficient was significant (β = .15, Study 2. As expected, the interaction coefficient
p < .05). To meet the second criterion for medi- significantly predicted self-esteem ( β = .24,
ation, the mediator must be significantly pre- p < .05) and perceived performance (β = .21,
dicted by the interaction between leadership p < .05). Using a similar hierarchical regression
efficacy and stereotype activation. Stereotype analysis as above, perceived performance
activation and leadership efficacy had significant significantly predicted self-esteem in Step 2
interactive effects on perceived performance (β = .57, p < .001). When the interaction term
(β = .23, p < .01). was entered into the equation with perceived
To test the remaining criteria for mediation, a performance it no longer significantly predicted
hierarchical regression analysis was performed. self-esteem (β = .12, p > .15), however, perceived
Predicting domain identification, the main performance still significantly predicted self-
effects for leadership efficacy and stereotype esteem (β = .54, p < .001).
activation were entered on Step 1, the mediator, Finally, to test whether the effect was different
perceived performance, was entered on Step 2, from zero we calculated the Goodman (I)
606

Table 1. Summary of hierarchical regression analyses testing perceived performance as a mediator of the
interaction between leadership efficacy and stereotype activation on domain identification
(studies 1 and 2 combined), self-esteem, and depressed affect (Study 2 only)
Mediator:
With no mediator Perceived performance
Domain identification
Step 1
Main effectsa R2 47.8% 47.8%
Step 2
Mediator β – .32***
∆R 2 – 8.3%
Step 3
Efficacy × Stereotype β .15* .08
∆R 2 2.3% 0.6%
Self-esteem
Step 1
Main effects a R2 20.5% 20.5%
Step 2
∆R 2 – 24.6%
Step 3
∆R 2 5.6% 1.4%
Depressed affect
Step 1
Main effectsa R2 19.5% 19.5%
Step 2
Mediator β – –.54***
∆R 2 – 22.1%
Step 3
∆R 2 5.1% 1.2%
Mediator β – –.51***
a
Block includes leadership efficacy and stereotype activation.
* p < .05; ** p < .01; *** p < .001.
version of the Sobel test statistic which revealed perceived performance was tested (see Table 1).
a marginal level of significance (Z = 1.89, p < .06). First, as expected, the interaction coefficient
Taken together, these mediational analyses between self-efficacy and stereotype activation
indicate that perceived performance mediates significantly predicted depressed affect (β = .23,
the interaction between stereotype activation p < .05). Using hierarchical regression, perceived
and leadership efficacy on self-esteem (see performance significantly predicted depressed
Figure 5). affect in Step 2, after the main effects were
entered into the equation (β = –.54, p < .001).
Depressed affect Finally, using data from Study 2, When the interaction term was entered in Step 3
the hypothesis that the interaction between it no longer significantly predicted depressed
stereotype activation and leadership efficacy affect (β = .11, p > .20) and perceived perform-
on depressed affect would be mediated by ance still significantly predicted depressed affect
607

*p < .05; **p < .01; ***p < .001.

Figure 5. Perceived performance as a mediator of the interactive effects of leadership efficacy and stereotype
activation on domain identification, self-esteem (Study 2), and depressed affect (Study 2).
(β = –.51, p < .001). Finally, the Goodman (I) interaction between leadership efficacy and
Sobel test statistic was marginally significant stereotype activation on depressed affect (see
(Z = 1.84, p < .07). Thus, perceived performance Figure 5).
also appears to be a successful mediator of the
608

General discussion for changes in domain identification and

psychological well-being resulting from stereo-
Substantial evidence highlights the incongruity type activation: self-perception. That is, the
between the female gender stereotype and the present studies tested the hypotheses that the
leadership role (Eagly & Karau, 2002). This interaction between leadership efficacy and
research examined how this incongruity affects stereotype activation on leadership domain
female leaders. More specifically, this research identification and psychological well-being
tested the moderating role of leadership efficacy would be mediated by the leader’s perception
on women leaders’ responses to the negative of her performance on the leadership task. As
stereotype regarding female leaders. In this hypothesized, regression analyses indicated
research, it was predicted that when blatantly that the effect of stereotype activation on high
presented with the gender leader stereotype and low efficacy leaders’ domain identification
high efficacy leaders would exhibit reactance to was mediated by the leaders’ perceived perform-
the stereotype; these hypotheses were strongly ance. Moreover, perceived performance also
supported across the two studies. appears to be a mechanism through which the
Leadership tasks often lack clear, immediate leaders’ self-reported well-being is influenced.
markers of success. On such tasks, the leaders’ These hypotheses were tested on both measures
immediate subjective sense of how well they of well-being in Study 2 and were supported.
performed becomes quite important. The It is important to note that reverse causal effects
hypothesis that stereotype activation would for the mediation results cannot completely be
serve to increase high efficacy leaders’ perceived ruled out in the present research. That is, it may
performance was supported. In addition, high be the case that domain identification and well-
efficacy leaders were rated as performing better being ‘cause’ perceived performance. However,
in the stereotype activation condition compared this alternative interpretation is less plausible
to the control condition. Domain identification when considering the theoretical work on self-
was another outcome examined in these studies. perception theory, which describes the import-
The seminal work on stereotype threat proposed ance of self-perception on inferences regarding
that one approach individuals may take to cope one’s traits and attitudes (Bem, 1967), and the
with the threat is to protectively disidentify with work on the sources of self-esteem which dem-
the domain (Steele & Aronson, 1995). Our re- onstrates the important impact of perceived
actance prediction that high efficacy women personal competence and living up to one’s
would evince heightened identification with personal standards of merit on self-esteem
the domain was supported. (Bandura, 1997). Our causal interpretation is
Finally, previous research highlights the also consistent with the theoretical and empir-
deleterious effects that negative stereotypes ical work showing that attributing successful
can have on the psychological well-being of the performance outcomes to internal causes, such
targets of the stereotypes (Swim et al., 2001). as ability or effort, raises self-esteem (Hoyt,
The present research tested the hypothesis that Aguilar, Kaiser, Blascovich, & Lee, in press;
high efficacy targets would show an increase in Weiner, 1985). Additionally, empirical research
well-being in the face of stereotype activation. testing alternative causal models between
This hypothesis was tested and supported in actual and perceived performance and self-
Study 2 and both self-esteem and depressed esteem also supports our causal interpretation
affect were taken as indices of psychological well- (Bohrnstedt & Felson, 1983). In sum, perceived
being. It appears that stereotype activation has a performance appears to play an important
positive effect on the self-esteem and depressed mediating role in the domain identification and
affect of high efficacy leaders. well-being effects of stereotype activation on
A second objective of this research was to women leaders; however, due to the correlational
explore one possible mechanism accounting nature of this research a better understanding
609

of the causal relationships between perceived activation to either have no effect or to be

performance, domain identification, and well- damaging to the targets of that stereotype. The
being merits further investigation. present research supports recent research by
underscoring the counterintuitive notion that
Implications stereotype activation can actually elicit positive
Theoretical This research expands our under- responses from targets of the stereotypes (Kray
standing of the effects of stereotypes beyond the et al., 2004). Finally, this work highlights the
academic domain to the domain of leadership. role of perceived performance in influencing
While investigations into the role stereotype the leaders’ identification with the leadership
activation plays within domains other than domain and their psychological well-being.
academia, such as decision task performance Perceptions of one’s performance seem to be
(Leyens, Désert, Croset, Darcis, 2000), has begun, an important source of immediate feedback
understanding the effects of stereotype activa- that appears to play a role in domain identifi-
tion across domains is in its infancy. In addition, cation and psychological well-being.
these experiments examined rated perform-
ance as well as self-concept outcomes deemed Applied value This line of research not only
important in these situations: perceived perform- provides important contributions to social psy-
ance, domain identification, and psychological chological theory, but it also contributes to a prac-
well-being. Thus, the results from the present tical understanding of the unique challenges
research highlight the important role that stereo- that women face in the realm of leadership.
type activation plays in influencing theoretically This research points to the resilience of women
relevant self-concept variables. who are highly confident in their leadership
Next, a comprehensive understanding of the abilities and are able to rise to the occasion in the
nature of stereotype activation should include face of negative stereotypes. When confronted
an insight into important moderator variables. with the stereotype these women perceive they
Previous research has documented both situ- perform better, are rated as performing better,
ational and individual difference moderators of become more identified with the domain of
stereotype threat. For example, task difficulty leadership, and have greater psychological
and frustration, test diagnosticity, and stereotype well-being.
relevance are situational factors that moderate While we cannot ascertain causality with the
reactions to stereotype threat, whereas domain approach taken in this research, these findings
identification, stigma consciousness, and iden- reveal that efficacy clearly moderates reactions
tification with the group to which the negative to stereotype activation. Thus, in a more applied
stereotype applies are individual difference vein, knowing the important moderating role
moderators (Aronson et al., 1999; Blascovich, of efficacy helps to identify those who may be
Spencer, Quinn, & Steele, 2001; Pinel, 1999; more or less adversely affected by stereotype
Schmader, 2002; Spencer, Iserman, Davies, & activation. Consequently, interventions designed
Quinn, 2001; Steele & Aronson, 1995). The cur- to foster more positive responses to stereotype
rent research demonstrates another important activation can be focused on those most likely
individual difference variable that moderates to benefit: those with low self-efficacy. Recent
reactions to stereotype activation: self-efficacy. research has suggested that making group
While this work has provided a deeper under- achievements salient (focusing on successful
standing about domain generalizability, out- women mathematicians) helps alleviate women’s
comes, and moderators of stereotype activation mathematical stereotype threat (McIntyre,
effects, an equally important theoretical ad- Paulson, & Lord, 2003). Thus, presenting low
vancement of this work includes understand- efficacy women with examples of successful
ing that, at times, stereotype activation can have women leaders may be effective in encouraging
a positive effect on performance. Until recently, more positive responses to stereotype activation
research has primarily shown stereotype in the leadership domain.
610

Finally, this research examined leadership activation has a particularly pernicious effect
within an immersive virtual environment. Given on the targets of the stereotypes. The impact of
the expansion of the virtual workplace, these more subtle stereotype activation on high and
findings are directly relevant to current organ- low efficacy leaders is an important empirical
izational trends. Additionally, the success of this question that merits further investigation.
research paradigm lends credibility to the use of Other differences between this research and
IVET in the study of leadership. Indeed, IVET stereotype threat research involve the differing
is a promising tool for the study of leadership domains, tasks, and outcome measures. First, the
by researchers as well as organizations. stereotype relevant domain often examined in
the stereotype threat literature is an academic
Limitations and future directions domain. The academic and leadership domains
It is important to point out that the current differ on a number of factors including the
findings are somewhat inconsistent with find- extent to which there are objective measures of
ings from the stereotype threat literature. In success. It may be that the effects of stereotype
most of the stereotype threat research, the activation are different in the domain of leader-
participants selected for the studies are those ship than the domain of academia, or on tasks
highly identified with the academic domain that either lack or have objective performance
under study. Indeed, studies designed to test the markers. Indeed the leadership task may not
role of domain identification in the stereotype have engendered the level of frustration often
threat paradigm generally reveal that stereotype encountered in academic tasks, and task frustra-
threat effects occur only for those who are highly tion has been implicated as a key component of
identified with the domain (Aronson et al., 1999). the stereotype threat process (Steele & Aronson,
The results of the present research indicate that 1995). Additionally, the outcomes examined in
stereotype activation had positive effects on this research differ from the classic stereotype
those who have high efficacy in their leadership threat outcomes. That is, the stereotype threat
abilities. Because self-efficacy and domain iden- literature often examines objective performance
tification are related constructs, the results from on an academic exam whereas the outcomes
the present research and the stereotype threat examined here were subjective performance
literature are somewhat at odds. ratings and self-concept outcomes associated
There are a few factors that may help explain with task performance. Stereotypes may not
the contradictory findings; these factors should affect these varying responses in the same
be further investigated. First and foremost, manner. A full examination of the parameters
stereotypes were activated blatantly in this re- of the effect will likely elucidate the apparent
search. The relative explicit vs. implicit manner contradiction in findings.
in which the stereotype is primed has been
shown to play an important role in responses Determining the causal role of efficacy Future
to these stereotypes (Kray et al., 2001). For ex- research should attempt to determine the causal
ample, recent research revealed that women role of leadership efficacy in the observed
primed with implicit gender stereotypes showed effects. According to Bandura (1997), there are
less of a preference for a leadership role com- four sources of efficacy information: enactive
pared to those not primed with the stereotype mastery, vicarious experiences, verbal persuasion,
(Davies, Spencer, & Quinn, 2002). Stoddard and physiological and affective states. Future
et al. (2003) also found that subtle stereotypes research should take a well-rounded approach
evoked assimilation patterns such that women to manipulating efficacy by focusing on these
were less likely to desire a leadership role; how- principal sources. Perhaps a promising approach
ever, they also found that blatant stereotypes to understanding the causal role of efficacy in
resulted in women demonstrating an increased the observed effects would be to develop a com-
desire to assume a leadership position. Thus, prehensive efficacy-training program, includ-
previous research indicates that subtle stereotype ing methods from each of the four sources of
611

efficacy information. Examining and comparing (a stereoscopic display with dual 680 horizontal
responses to stereotype activation of low efficacy by 480 vertical resolution LCD panels that
women who have and have not had the training refresh at 72 Hz). The optics of this display
may provide us with insight into the causal role presented a visual stimulus subtending
approximately 50 degrees horizontally by
of efficacy in reactions to stereotype activation.
38 degrees vertically. Perspectively correct
Additionally, the role of efficacy in reactions to stereoscopic images were rendered by a 450
stereotype activation in other domains should MHz Pentium III dual processor computer
be evaluated. with an Evans & Sutherland Tornado 3000
dual pipe graphics card, and these images were
updated at an average frame rate of 36 Hz. The
Conclusions simulated viewpoint was continually updated
by the participants’ head movements. The
The incongruity between the leadership role
orientation of the participant’s head was tracked
and the female gender role has important by a three-axis orientation sensing system
implications for women leaders. The current (Intersense IS300, update rate of 150 Hz). The
research highlights the role of leadership system latency, or the amount of delay between
efficacy in responses to negative stereotypes a participant’s head motion and the resulting
in the domain of leadership. As demonstrated concomitant update in the HMD’s visual display,
here, stereotype activation can actually be asso- was 65 ms maximum.
ciated with beneficial responses—but only for 4. During debriefing only four participants
women who have high self-efficacy for leader- indicated suspicion that there were no other
ship. Currently, most research is focused on participants. Across both studies, analyses with
and without suspicious participants yield similar
examining the debilitating effects that negative
results.
stereotypes have on the targets of these stereo- 5. These statistics were accurate at the time the
types. Hopefully this work will open up an experiment was conducted.
avenue of investigation into the more beneficial 6. Importantly, reverse causal effects cannot
responses to these stereotypes. A better under- completely be ruled out. That is, oftentimes
standing of high efficacy leaders’ responses to in mediational analyses if the mediator and
stereotype activation may provide us with tools the outcome variable are interchanged, the
to counter the more detrimental responses of outcome seems to ‘cause’ the mediator (Kenny,
those lacking efficacy. Kashy, & Bolger, 1998). Accordingly, in the
present research interchanging perceived
performance with the outcome variables yielded
Notes significant mediational effects. However, the
plausibility of reverse causation is weakened by
1. Researchers have adopted varying labels for the experimental design in which the mediator
the distinction between the descriptive and was measured temporally prior to the outcome
prescriptive aspects of gender stereotypes. Eagly variables. The analyses in this manuscript focus
and Karau (2002) use role terminology, referring on the causal paths based on a priori theoretical
to the descriptive and injunctive aspects of predictions.
gender roles, whereas others use stereotype
terminology, referring to descriptive stereotypes,
or gender stereotypes, and prescriptive
Acknowledgements
stereotypes (Burgess & Borgida, 1999). This This work was supported in part by NSF ITR
paper makes no distinction between the terms Award #0205740 to the second author. We
and uses them interchangeably. acknowledge and thank Lauren Aguilar, Donovan
2. In this manuscript, priming the stereotype refers Bean, Missy Clayton, Michelle Dennis, Sarah
to the general operationalization of stereotype Estrada, Lauren Gase, Sarah Haskell, Diana Hill,
activation. Annie Lamson, Kevin Lee, Elizabeth LeMoine,
3. Technology specifications: The head mounted Cari Nicholson, and Peter Westphalen for their
display was a Virtual Research V8 HMD superb contributions to this research.
612

References Bohrnstedt, G. W., & Felson, R. (1983).

Explaining the relations among children’s
Arkkelin, D., & Simmons, R. (1985). The ‘good actual and perceived performances and self-
manager’: Sex-typed, androgynous, or likable? esteem: A comparison of several causal models.
Sex Roles, 12, 1187–1198. Journal of Personality and Social Psychology,
Aronson, J., Lustina, M. J., Good, C., Keough, 45, 43–56.
K., Steele, C. M., & Brown, J. (1999). When Boldry, J., Wood, W., & Kashy, D. A. (2001). Gender
White men can’t do math: Necessary and stereotypes and the evaluation of men and
sufficient factors in stereotype threat. Journal of women in military training. Journal of Social
Experimental Social Psychology, 35, 29–46. Issues, 57, 689–705.
Aronson, J., Quinn, D. M., & Spencer, S. J. Brehm, J. W. (1966). A theory of psychological
(1998). Stereotype threat and the academic reactance. New York: Academic Press.
underperformance of minorities and women. Broverman, I.K., Vogel, S.R., Broverman, D.M.,
In J. K. Swim & C. Stangor (Eds.), Prejudice: Clarkson, F.E., & Rosenkrantz, P.S. (1972).
The target’s perspective (pp. 83–103). San Diego, Sex-role stereotypes: A current reappraisal.
CA: Academic Press, Inc. Journal of Social Issues, 28, 59–78.
Bandura, A. (1982). Self-efficacy mechanism in Bureau of Labor Statistics (2006). Women in the labor
human agency. American Psychologist, 37, 122–147. force: A databook. Retrieved 2 December 2006
Bandura, A. (1986). Social foundations of thought and from: http://www.bls.gov/cps/wlf-databook–
action. Englewood Cliffs, NJ: Prentice-Hall. 2006.pdf
Bandura, A. (1997). Self-efficacy: The exercise of Burgess, D., & Borgida, E. (1999). Who women
control. New York: W.H. Freeman. are, who women should be: Descriptive
Bandura, A., & Cervone, D. (1983). Self-evaluative and prescriptive gender stereotyping in sex
and self-efficacy mechanisms governing the discrimination. Psychology, Public Policy, & Law,
motivational effects of goal systems. Journal of 5, 665–692.
Personality and Social Psychology, 45, 1017–1028. Campbell, N.K., & Hackett, G. (1986). The effects
Bandura, A., & Wood, R. (1989). Effect of of mathematics task performance on math self-
perceived controllability and performance efficacy and task interest. Journal of Vocational
standards on self-regulation of complex Behavior, 28, 149–162.
decision-making. Journal of Personality and Social Carli, L. L. (2001). Gender and social influence.
Psychology, 56, 805–814. Journal of Social Issues, 57, 725–741.
Barling, J., & Beattie, R. (1983). Self-efficacy beliefs Carli, L.L., & Eagly, A. H. (2001). Gender,
and sales performance. Journal of Organizational hierarchy, and leadership: An introduction.
Behavior Management, 5, 41–51. Journal of Social Issues, 57, 629–636.
Baron, R. M., & Kenny, D. A. (1986). The Catalyst (2002). Catalyst census of women corporate
moderator–mediator variable distinction in officers and top earners. New York: Catalyst.
social psychological research: Conceptual, Chemers, M. M., Watson, C. B., & May, S. (2000).
strategic and statistical considerations. Journal of Dispositional affect and leadership effectiveness:
Personality and Social Psychology, 51, 1173–1182. A comparison of self-esteem, optimism and
Bem, D. (1967). Self-perception: An alternative efficacy. Personality and Social Psychology Bulletin,
interpretation of cognitive dissonance 26, 267–277.
phenomena. Psychological Review, 74, 183–200. Davies, P.G., Spencer, S. J., & Quinn, D. M.
Bergeron, D.M., Block, C.J., & Echtenkamp, B.A. (2002). Consuming images: How television
(2006). Disabling the able: Stereotype threat and commercials that elicit stereotype threat can
women’s work performance. Human Performance, restrain women academically and professionally.
19, 133–158. Personality and Social Psychology Bulletin, 28,
Blascovich, J., Loomis, J., Beall, A., Swinth, K., Hoyt, 1615–1628.
C., & Bailenson, J. (2002). Immersive virtual Davies, P. G., Spencer, S. J., & Steele, C. M. (2005).
environment technology as a methodological Clearing the air: Identity safety moderates
tool for social psychology. Psychological Inquiry, the effects of stereotype threat on women’s
13, 103–125 leadership aspirations. Journal of Personality and
Blascovich, J., Spencer, S. J., Quinn, D., & Steele, Social Psychology, 88, 276–287.
C. (2001). African Americans and high blood Deaux, K., & Lewis, L. L. (1983). Components
pressure: The role of stereotype threat. of gender stereotypes. Psychological Documents,
Psychological Science, 12, 225–229. 13(2), MS. 2583 25.
613

Deaux, K., & Lewis, L. L. (1984). Structure of Hill, T., Smith, N.D., & Mann, M.F. (1987). Role of
gender stereotypes: Inter-relationships among efficacy expectations in predicting the decision
components and gender label. Journal of to use advanced technologies. Journal of Applied
Personality and Social Psychology, 46, 991–1004. Psychology, 72, 307–314.
Dodge, K.A., Gilroy, F. D., & Fenzel, L. M. (1995). Hoyt, C., Aguilar, L., Kaiser, C., Blascovich, J.,
Requisite management characteristics revisited: & Lee, K. (in press). The self-protective and
Two decades later. Journal of Social Behavior & undermining effects of attributional ambiguity.
Personality, 10, 253–264. Journal of Experimental Social Psychology.
Eagly, A. H. (1987). Sex differences in social behavior: Inzlicht, M., & Ben-Zeev, T. (2000). A threatening
A social-role interpretation. Hillsdale, NJ: Erlbaum. intellectual environment: Why females are
Eagly, A. H., & Karau, S. (2002). Role congruity susceptible to experiencing problem-solving
theory of prejudice toward female leaders. deficits in the presence of males. Psychological
Psychological Review, 109, 573–598. Science, 11, 365–371.
Eagly, A. H., Makhijani, M. G., & Klonsky, B. G. Jex, S. M., & Bliese, P. D. (1999). Efficacy beliefs
(1992). Gender and the evaluation of leaders: A as a moderator of the impact of work-related
meta-analysis. Psychological Bulletin, 111, 3–22. stressors: A multilevel study. Journal of Applied
Eagly, A. H., Wood, W., & Diekman, A. B. (2000). Psychology, 84, 349–361.
Social role theory of sex differences and Joyce, A. (2006, 15 August). Fortune 500 gains
similarities: A current appraisal. In T. Eckes & female CEO in new PepsiCo chief. The
H. M. Trautner (Eds.), The developmental social Washington Post, pp. D01.
psychology of gender (pp. 123–174). Mahwah, Kenny, D. A., Kashy, D. A., & Bolger, N. (1998).
NJ: Erlbaum. Data analysis in social psychology. In D. Gilbert,
Eckes, T. (1994). Explorations in gender cognition: S. Fiske, & G. Lindzey (Eds.), The handbook of
Content and structure of female and male social psychology (Vol. 1, 4th ed., pp. 233–265).
subtypes. Social Cognition, 12, 37–60. Boston, MA: McGraw-Hill.
Glick, P., & Fiske, S. T. (1996). The Ambivalent Kray, L., Reb, J., Galinsky, A. & Thompson, L.
Sexism Inventory: Differentiating hostile and (2004). Stereotype reactance at the bargaining
benevolent sexism. Journal of Personality and table: The effect of stereotype activation and
Social Psychology, 70, 491–512. power on claiming and creating value.
Glick, P., & Fiske, S. T. (1999). Sexism and other Personality and Social Psychology Bulletin,
‘isms’: Independence, status, and the ambivalent 30, 399–411.
content of stereotypes. In W. B. Swann, Jr. & Kray, L. J., Thompson, L., & Galinsky, A. (2001).
J. H. Langlois (Eds.), Sexism and stereotypes in Battle of the sexes: Gender stereotype
modern society: The gender science of Janet Taylor confirmation and reactance in negotiations.
Spence (pp. 193–221).Washington, D.C.: Journal of Personality and Social Psychology, 80,
American Psychological Association. 942–958.
Heatherton, T.F., & Polivy, J. (1991). Development Lazarus, R. S., & Folkman, S. (1984). Stress, coping,
and validation of a scale for measuring state self- and adaptation. New York: Springer.
esteem. Journal of Personality and Social Psychology, Lent, R.W., Brown, S.D., & Larkin, K.C. (1987).
60, 895–910. Comparison of three theoretically derived
Heilman, M. E. (2001). Description and variables in predicting career and academic
prescription: How gender stereotypes prevent behavior: Self-efficacy, interest congruence,
women’s ascent up the organizational ladder. and consequence thinking. Journal of Counseling
Journal of Social Issues, 57, 657–674. Psychology, 34, 293–298.
Heilman, M. E., Block, C. J., & Martell, R. F.(1995). Leyens, J., Désert, M., Croizet, J., & Darcis, C.
Sex stereotypes: Do they influence perceptions (2000). Stereotype threat: Are lower status
of managers? Journal of Social Behavior and and history of stigmatization preconditions of
Personality, 10, 237–252. stereotype threat? Personality & Social Psychology
Heilman, M. E., Block, C. J., Martell, R. F., & Bulletin, 26, 1189–1199.
Simon, M. C. (1989). Has anything changed? Lubin, B., Zuckerman, M., & Woodward, L. (1985).
Current characterizations of men, women Bibilography for the Multiple Affect Adjective
and managers. Journal of Applied Psychology, 74, Check List. San Diego, CA: Educational and
935–942. Industrial Testing Service.
614

Major, B., Spencer, S., Schmader, T., Wolfe, C., Schein, V. E. (1973). The relationship between
& Crocker, J. (1998). Coping with negative sex role stereotypes and requisite management
stereotypes about intellectual performance: characteristics. Journal of Applied Psychology,
The role of psychological disengagement. 57, 95–100.
Personality & Social Psychology Bulletin, 24, 34–50. Schein, V. E. (2001). A global look at psychological
Martell, R. F., Parker, C., Emrich, C. G., & barriers to women’s progress in management.
Crawford, M. S. (1998). Sex stereotyping in the Journal of Social Issues, 57, 675–688.
executive suite: ‘Much ado about something’. Schmader, T. (2002). Gender identification
Journal of Social Behavior and Personality, 13, moderates stereotype threat effects on women’s
127–138. math performance. Journal of Experimental Social
McIntyre, R. B., Paulson, R. M., & Lord, C. G. Psychology, 38, 194–201.
(2003). Alleviating women’s mathematics Sczesny, S. (2003). A closer look beneath the
stereotype threat through salience of group surface: Various facets of the think-manager-
achievements. Journal of Experimental Social think-male stereotype. Sex Roles, 49, 353–363.
Psychology, 39, 83–90. Sekaquaptewa, D., & Thompson, M. (2003). Solo
Murphy, S. E. (1992). The contribution of leadership status, stereotype threat, and performance
experience and self-efficacy to group performance under expectancies: Their effects on women’s
evaluation apprehension. Unpublished doctoral performance. Journal of Experimental Social
dissertation, University of Washington, Seattle. Psychology, 39, 68–74.
Murphy, S. E., Chemers, M. M., Kohles, J., & Sobel, M. E. (1982). Asymptotic intervals for
Macaulay, J. L. (2004). The contribution of indirect effects in structural equations models.
leadership self-efficacy to performance under stress: In S. Leinhart (Ed.), Sociological methodology
An extension of cognitive resources theory. 1982 (pp. 290–312). San Francisco:
Unpublished manuscript. Jossey-Bass.
Murphy, S. E. (2002). Leader self-regulation: Spencer, S. J., Iserman, E., Davies, P. G., & Quinn,
The role of self-efficacy and multiple D. M. (2001). Suppression of doubts, anxiety, and
intelligences. In R.E. Riggio & S. E. Murphy stereotypes as a mediator of the effect of stereotype
(Eds.), Multiple intelligences and leadership threat on women’s math performance. Unpublished
(pp. 163–186). Mahwah, NJ: Lawrence manuscript, University of Waterloo, Canada.
Erlbaum and Associates. Steele, C. M., & Aronson, J. (1995). Stereotype
Pinel, E. C. (1999). Stigma consciousness: The threat and the intellectual test performance
psychological legacy of social stereotypes. of African Americans. Journal of Personality and
Journal of Personality and Social Psychology, 76, Social Psychology, 69, 797–811.
114–128. Stoddard, T., Kliengklom, T., & Ben-Zeev, T.
Powell, G. N., & Butterfield, D. A. (1979). The (2003, February). Stereotype threat, assimilation,
‘good manager’: Masculine or androgynous? and contrast effects, and subtlety of priming, or: ‘You
Academy of Management Journal, 22, 395–403. say “Bitch” like it’s a bad thing’. Paper presented
Powell, G. N., & Butterfield, D. A. (1984). If at the annual Society for Personality and Social
‘good managers’ are masculine, what are ‘bad Psychology conference, Los Angeles, CA.
managers’? Sex Roles, 10, 477–484. Stumpf, S. A., Brief, A. P., & Hartman, K. (1987).
Powell, G. N., & Butterfield, D. A. (1989). The Self-efficacy expectations and coping with
‘good manager’: Did androgyny fare better in career-related events. Journal of Vocational
the 1980s? Group and Organization Studies, 14, Behavior, 31, 91–108.
216–233. Swim, J. K., Hyers, L. L., Cohen, L. L., & Ferguson,
Ridgeway, C. L. (2001). Gender, status, and M. J. (2001). Everyday sexism: Evidence for its
leadership. Journal of Social Issues, 57, 637–655. incidence, nature, and psychological impact
Rosenwasser, S. M., & Dean, N. G. (1989). Gender from three daily diary studies. Special Issue:
role and political office: Effects of perceived Stigma: An insider’s perspective. Journal of Social
masculinity/femininity of candidate and Issues, 57, 31–53.
political office. Psychology of Women Quarterly, 13, Taylor, M.S., Locke, E.A., Lee, C., & Gist, M.E.
77–85. (1984). Type A behavior and faculty research
Rudman, L. A., & Glick, P. (2001). Prescriptive productivity: What are the mechanisms?
gender stereotypes and backlash toward agentic Organizational Behavior and Human Decision
women. Journal of Social Issues, 57, 743–762. Processes, 34, 402–418.
615

Watson, C. B., Chemers, M. M., & Preiser, N. Biographical Notes

(1996, June). Collective efficacy: A multi-level
analysis. Presented at the annual meeting of crystal hoyt completed her doctorate in social
the American Psychological Society, psychology at the University of California, Santa
San Francisco, CA. Barbara. She is currently an assistant professor
Weiner, B. (1985). An attributional theory at the Jepson School of Leadership Studies at
of achievement motivation and emotion. the University of Richmond. Her curricular
Psychological Review, 92, 548–573. and research interests include social behavior,
White, J., & Gardner, W. (2003, February). leadership and group dynamics, the effects
Confidence as a moderator of the effect of solo of stereotypes and discrimination on women
status on performance in a workgroup. Paper and minority leaders, the role of confidence in
presented at the annual Society for Personality shaping group leadership, leader perception, and
and Social Psychology conference, Los Angeles, new methodological tools for social scientists.
CA.
Williams, J. E, & Best, D. L. (1990). Sex and psyche: jim blascovich earned his PhD in social
Gender and self viewed cross-culturally. Thousand psychology at the University of Nevada, Reno.
Oaks, CA: Sage Publications, Inc. He is currently professor of psychology at the
Wood, W., Christensen, P. N., Hebl, M. R., & University of California, Santa Barbara. He
Rothgerber, H. (1997). Conformity to sex-typed co-directs the Research Center for Virtual
norms, affects, and the self-concept. Journal of Environments and Behavior. His main
Personality and Social Psychology, 73, 523–535. research interests include challenge and
threat motivation, and the use of immersive
Paper received 26 November 2005; revised version virtual envirnment technology (IVET) as a
accepted 4 January 2007. methodological tool in social psychology.
616

595 Full

Uploaded by

Copyright:

Available Formats

595 Full

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

595 Full

Uploaded by

Copyright:

Available Formats

Group Processes & Intergroup

Leadership Efficacy and Women Leaders' Responses to Stereotype Activation

The online version of this article can be found at:

Email Alerts: http://gpi.sagepub.com/cgi/alerts

>> Version of Record - Nov 29, 2007

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Leadership Efficacy and

The role of leadership efficacy in women’s reactance responses to stereotype-based leadership

keywords gender, leadership, reactance, self-efficacy, small groups, stereotypes

Organizational leaders make important Karau (2002) recently proposed an explanation

Copyright © 2007 SAGE Publications

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

The impact of stereotype activation to possess sufficient power in order to react

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

assume a more emotion-focused approach. perseverance, resilience in the face of adversity,

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Mediational hypotheses a paradigm designed to test the hypotheses in

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

environments and their contents as if they are

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Participant selection: Self-efficacy for leadership Stereotype activation manipulation To prime

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

higher levels of perceived performance (M = 1.17,

Figure 3. The impact of leadership efficacy and

Rated performance Unfortunately, a few of the

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

identification in Study 2 than Study 1, and low

Moderation hypotheses: Self-esteem and

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

*p < .05; **p < .01; ***p < .001.

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

General discussion for changes in domain identification and

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

of the causal relationships between perceived activation to either have no effect or to be

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

References Bohrnstedt, G. W., & Felson, R. (1983).

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

Watson, C. B., Chemers, M. M., & Preiser, N. Biographical Notes

Downloaded from gpi.sagepub.com at CENTIC on November 19, 2012

You might also like

p < .05; p < .01; p < .001.