PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN Rudman, Kilianski / FEMALE AUTHORITY ATTITUDES
Implicit and Explicit Attitudes
Toward Female Authority
Laurie A. Rudman
Stephen E. Kilianski
Attitudes toward female authority and their relationship to gen
der beliefs were examined using implicit and explicit measures of
each. Implicit attitudes covaried with implicit gender authority
beliefs (i.e., linking men to high-authority and women to
low-authority roles). Explicit attitudes covaried with explicit
gender authority beliefs, feminist identification, and hostile sex-
ism. Thus, gender authority beliefs may influence both conscious
and unconscious prejudice against female authorities.
Although women showed less explicit prejudice than did men,
their implicit attitudes were similarly negative. Finally, the rela-
tionship found between two different response latency methods (a
priming task for attitudes, a categorization task for beliefs) sup-
ports the assumption that implicit measures assess similar con-
structs (i.e., automatic associations in long-term memory).
When a 747 hits turbulence and the pilot’s assur
ances waft over the intercom, passengers may be more
soothed by a baritone than a soprano voice. Patients
undergoing operations may be relieved to see hirsute
forearms protruding from the surgeon’s gloves. There is
evidence that men and women alike prefer having a man
in charge. A 1996 international Gallup Poll found that
the majority of respondents favored male as opposed to
female bosses (Gallup, 1996; see also Rubner, 1991). Sur
prisingly, women expressed this preference more often
than did men. In the laboratory, too, men are preferred
to women as experts and leaders (e.g., Eagly & Karau,
1991; Watson, 1988; Wright, 1976). Consistent with this
bias, women continue to exercise less authority in their
jobs than do men (Moore & Shackman, 1996; Wright &
Baxter, 1995). As a result, women are promoted and
compensated at rates significantly less than those of men
(Lyness & Thompson, 1997; Reskin & Ross, 1992;
Sonnert & Holton, 1996). Gender inequity vis-à-vis
authority is clearly costly to women, then. The present
research used both implicit (i.e., automatic) and explicit
(i.e., controlled) methods to investigate why men are
preferred to women as leaders: What is it about men that
comforts people when they are in charge?
Authority comes in many forms, several of which,
including expert, legitimate, and coercive authority, are
the purview of men (Johnson, 1976). The gender gap in
authority reflects chronic power differences between
men and women. This disparity, prevalent throughout
history and across cultures, may stem from traditional
labor divisions, in which men and women have tradition-
ally been assigned occupational and domestic roles,
respectively (Eagly & Wood, 1991). Despite the fact that
women now represent half of all workers, perceptions
that social roles differ for men and women may be intact
(Carli & Eagly, in press). The gender role hypothesis is
derived from social role theory’s emphasis on traditional
labor divisions as a structural cause of gender inequities
(Eagly, 1987). To the extent that individuals associate
men with career and women with domestic roles, they
may view female authorities as violating traditional gen
der role assignments (e.g., family values).
Although social roles have dramatically changed for
women, they continue to be underrepresented in leader
ship roles. The gender authority hypothesis posits that
labor divisions within the workplace signify different
status expectancies for men and women. If gender
operates as a cue to legitimacy, men may be accorded
more prestige simply by virtue of being male (Berger,
Fisek, Norman, & Zelditch, 1977). Furthermore, male
dominance in powerful social roles (e.g., politics, law,
Authors’ Note: This research was partially supported by Grant
SBE-9807970 from the National Science Foundation to the first author.
We thank Richard Ashmore and Tony Greenwald for helpful com
ments on an earlier version of this article. Correspondence may be ad
dressed to Laurie A. Rudman, Department of Psychology, Tillett Hall,
Rutgers University, 53 Avenue E, Piscataway, NJ 08854-8040; e-mail:
PSPB, Vol. 26 No. 11, November 2000 1315-1328
© 2000 by the Society for Personality and Social Psychology, Inc.
religion, and the military) has produced an implicit
male leader prototype (Banaji & Greenwald, 1995; Eagly,
1987; Forsythe, Heiney, & Wright, 1997; Vianello &
Siemienska, 1990). This prototype may be both cause
and effect of a generalized belief that men are superior
and thus deserve to control and receive more resources
than do women (Jost & Banaji, 1994). As a result, the
association between men and authority may be stronger
than the association between women and authority. If so,
then powerful women may be disliked for breaching an
expectancy that men are natural leaders.
Finally, the gender stereotype hypothesis posits that dif
ferent trait expectancies for men and women underlie
negative attitudes toward female authority. Because pow
erful roles have traditionally been filled by men, author
ity itself may be more associated with male characteristics
(e.g., agency) than female characteristics (e.g.,
communality). This produces “lack of fit” perceptions
between women and power (Eagly, 1987; Heilman, 1983;
Heilman, Block, Martell, & Simon, 1989). Because gen
der beliefs are highly prescriptive as well as descriptive
(Glick & Fiske, 1999), differential role, authority, and
trait associations also might influence prejudice against
female authorities on the basis that women should be
less career oriented, authority seeking, and agentic (i.e.,
leader-like) than men (Carli & Eagly, in press; Glick &
These hypotheses are not mutually exclusive. In fact,
gender role, authority, and trait beliefs may be interre-
lated (e.g., Eagly & Steffen, 1984). As a result, each con-
struct may covary with attitudes toward female authority.
However, the relative strength of these relationships has
not been investigated using either explicit or implicit
A related question is the extent to which prejudice
against female authorities might operate implicitly (i.e.,
below conscious awareness). In general, gender atti
tudes and belief systems are multifaceted (Ashmore, Del
Boca, & Bilder, 1995) and undoubtedly contain uncon
scious as well as conscious elements. However, little
attention has been paid to implicit gender attitudes. To
date, the primary focus has been on implicit gender stereo
types (e.g., Blair & Banaji, 1996; Banaji & Greenwald, 1995;
Rudman, Greenwald, & McGhee, 1999; see Greenwald &
Banaji, 1995, for a review). The present research uniquely
examined implicit versus explicit attitudes toward
female authority. To assess the former, we used a priming
task previously shown to be an effective measure of
implicit racism (Fazio, Jackson, Dunton, & Williams,
1995). To assess the latter, we used a self-report instru
ment (the Gender and Authority Measure), designed
and piloted prior to conducting this research (Rudman,
Overview of Research
Gender role (career vs. domestic), gender authority
(high vs. low), and gender trait (agency vs. communality)
beliefs were examined as correlates of attitudes toward
female authorities. Participants’ implicit and explicit (a)
attitudes toward female authority and (b) gender beliefs
were obtained. As a means of comparison, implicit atti
tudes toward high-authority male and low-authority male
and female targets also were measured.
To provide explicit measures of gender beliefs, we
used traditional rating scales (e.g., Rudman et al., 1999).
To assess implicit gender beliefs, we used the Implicit
Association Test (IAT) (Greenwald, McGhee, &
Schwartz, 1998). Previous research has supported the
IAT as a powerful and flexible measure of unconscious
attitudes and beliefs, including gender stereotypes
(Rudman et al., 1999; Rudman, Greenwald, Mellott, &
Schwartz, in press). As with priming measures, the IAT
assesses response latency and accuracy for judgments
designed to be facilitated (or slowed down) by automatic
associations. The advantage of these methods is that they
do not rely on respondents’ ability or willingness to report
their attitudes (Dovidio & Fazio, 1992; Greenwald &
Banaji, 1995). These methods were necessary to exam-
ine unconscious attitudes toward female authority (and
implicit gender beliefs as potential predictors). But they
were also prudent, because explicit gender measures
may provoke social desirability concerns on the part of
respondents (e.g., Swim, Aiken, Hall, & Hunter, 1995).
The use of two different response latency methods (a
priming measure for attitudes, the IAT for beliefs) is
unique to the present research, but theoretically, each
assesses automatic associations in long-term memory.
Although these associations are evaluative and semantic,
respectively, they should covary (Breckler, 1984; Eagly,
1987). If implicit beliefs are linked to implicit attitudes,
then the results would support the construct validity of
each method. This is particularly important because these
measures typically show weak relations with their
self-report counterparts (e.g., Fazio et al., 1995; Greenwald
et al., 1998). Although the lack of implicit-explicit
covariation supports the discriminant validity of uncon
scious constructs (Greenwald & Banaji, 1995), implicit
methods should themselves overlap. The present
research provides a test of this hypothesis.
We also examined individual differences in attitudes
toward female authority. Considerable evidence shows
that women’s self-reported attitudes are more egalitar
ian than are men’s (e.g., Ashmore et al., 1995; Eagly,
Makhijani, & Klonsky, 1992; Glick & Fiske, 1996; Swim
et al., 1995; Williams & Best, 1990). Therefore, men were
expected to show more explicit prejudice against female
authorities than were women. However, sex differences
in implicit prejudice may be less likely given that sex dif
1316 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
ferences in implicit gender stereotypes are rare (e.g.,
Banaji & Hardin, 1996; Blair & Banaji, 1996; cf. Rudman
et al., 1999). Finally, feminist identity and hostile sexism
(i.e., antifemale attitudes) (Glick & Fiske, 1996) have
shown positive and negative relations, respectively, with
self-reported attitudes toward women in nontraditional
roles (e.g., Forsythe et al., 1997; Glick, Diebold,
Bailey-Werner, & Zhu, 1997). They were therefore
included as potential correlates for both explicit and
implicit attitudes toward female authority.
Seventy-five volunteers from introductory psychology
courses participated in exchange for course credit. Data
from 6 volunteers were discarded due to incomplete
response latency data, leaving 69 participants (35
women, 34 men). Of these, 46% were White, 29% were
Asian, and 16% were African American (9% were
Gender and Authority Attitudes
Priming materials. Implicit attitudes were assessed with
a priming measure (Fazio et al., 1995). Twenty-four
black-and-white schematic drawings of (White) men and
women in various occupations served as primes. Six
primes each (n = 12 primes) showed men and women as
authorities, matched on gender and role (doctor, profes-
sor, police officer, boss, judge, and scientist). Six primes
each also showed men and women in low-authority roles
(n = 12 primes). Eight of the low-authority primes were
matched on gender and role (cook, nurse, model,
waiter/waitress); four were unmatched (male hair
dresser, male baker; female maid, ballerina). Figure 1
shows examples of high- and low-authority female
Twelve positive and 12 negative adjectives served as
the target words. The positive adjectives were clever, good,
competent, healthy, intelligent, loyal, likable, optimistic, pleas
ant, smart, honest, and responsible. The negative adjectives
were bitter, annoying, careless, cowardly, cynical, dishonest, for
getful, gloomy, harmful, selfish, snobbish, and bossy. These
were selected on the basis of norms provided by Williams
and Best (1990). The objective was to ensure assessment
of implicit attitudes rather than gender stereotypes.
Overall, the positive and negative adjectives differed in
valence, t(22) = 30.83, p < .001 (Ms = 628 vs. 373, respec
tively) but did not differ in gender association, t(22) =
.33, p = .75 (Ms = 495 and 503, respectively).
Procedure. The procedure followed that of Fazio et al.
(1995). Computerized instructions informed partici
pants that the task measured their ability to memorize
pictures while judging the meaning of words.
Participants’ task was to press a key labeled “good” or a
key labeled “bad” as quickly as possible to indicate the
valence of 24 adjectives (12 positive, 12 negative). The
procedure involved five phases. In Phase 1, participants
performed a baseline task of 48 trials in which they
responded to each adjective twice. The adjectives (ran
domly presented) were preceded by a row of asterisks
presented for 450 ms. Adjectives remained on the screen
until participants indicated their judgment. A 2.5-s inter
val separated each trial. The mean latency for the two tri
als involving each adjective served as that adjective’s
In Phase 2, participants attended to 12 schematic tar
gets portraying men and women in different occupa
tions (e.g., paramedic, telephone operator, typists),
ostensibly presented as stimuli for an upcoming memory
task. Phase 3 involved a recognition test that included
Rudman, Kilianski / FEMALE AUTHORITY ATTITUDES 1317
Figure 1 Examples of schematic women used as priming stimuli.
Phase 2’s stimulus pictures and 12 foils.
pressed separate keys for “old” and “new” pictures dur
ing this phase. The purpose of Phase 3 was to bolster the
cover story and thereby increase the likelihood that par
ticipants would attend to the primes during Phase 4.
Phase 4 involved the actual priming task, used to
assess implicit attitudes toward female authority. The
procedure followed Phase 1’s procedure but the row of
asterisks was replaced by the primes (i.e., schematic men
and women shown as high or low authorities). Comput
erized instructions informed participants that the mem
ory and word judgment tasks would be combined. They
were told to memorize the pictures preceding each
adjective while continuing to perform quickly and accu
rately on the judgment task. There were four blocks of
trials in Phase 4. Each block consisted of 24 trials, in
which each prime appeared once, followed by one of the
24 adjectives. Throughout the course of four blocks,
each prime was paired with 2 positive and 2 negative
adjectives (total trials = 96). Moreover, each prime type
was paired once with each positive and negative adjec
tive. Consistent with the cover story, Phase 5 involved a
recognition test in which participants distinguished the
24 primes from 24 foils. The data from Phase 5 indicated
that participants attended to the primes (M correct =
84%, SD = 6.33).
Self-report measure. Explicit attitudes were assessed by
the Gender and Authority Measure (GAM). The GAM
consists of 15 items on which respondents indicate pref-
erence for male versus female authorities in five areas of
social influence (i.e., legitimate, expert, reward, coer-
cive, and referent) (French & Raven, 1959). Respon-
dents express agreement with each item on a scale rang
ing from 1 (strongly disagree)to5(strongly agree).
Examples of items include “If I were on trial, I would pre
fer that the judge be a man” (legitimate); “For most col
lege courses, I prefer a male to a female professor”
(expert); “I would rather work for a man than a woman”
(reward); “In general, I would rather take orders from a
man than a woman” (coercive); and “The people I look
up to most are men” (referent). The GAM is scored by
averaging respondents’ agreement with the 15 items,
after reverse-scoring appropriate items. High scores
indicate preference for male versus female authorities.
In the present sample, the GAM showed adequate inter
nal consistency (α = .82). The appendix contains the
Gender Beliefs Assessment
IAT stimulus materials. Implicit gender beliefs were
assessed with three IATs (Greenwald et al., 1998). The
gender roles IAT used 44 stimulus words: 15 male names
(e.g., Brian, Kevin, Paul), 15 female names (e.g., Meg,
Karen, Ann), 7 career-meaning words (career, job, sal
ary, office, promotion, finances, and occupation), and 7
domestic-meaning words (domestic, family, marriage,
child care, cooking, kitchen, and shopping).
der authority IAT used the same male and female names
in addition to 6 high-status occupational roles (boss,
executive, expert, leader, authority, and supervisor) and
6 low-status occupational roles (assistant, secretary,
clerk, subordinate, aid, and helper). The gender stereo
type IAT used the same male and female names in addi
tion to 7 agentic words (individualistic, competitive,
independent, challenging, self-sufficient, autonomous,
and hierarchical) and 7 communal words (communal,
connected, commitment, together, kinship, supportive,
and interdependent). Because prior research showed
that the evaluative connotations of judgments affect sex
differences in implicit gender stereotypes (Rudman et
al., 1999), the stimuli for the gender stereotype IAT were
pretested to be matched on valence.
IAT procedure. The five steps of the IAT, described with
materials for the gender stereotype task, were as follows:
(a) Participants distinguished target concepts by press-
ing the right key for male names and the left key for
female names, (b) participants distinguished the attrib-
ute dimension by pressing the right key for agentic words
and the left key for communal words, (c) participants
responded to male names and agentic words with the
right key and female names and communal words with
the left key (combined categorization task, abbreviated
as male + agency), (d) participants repeated Step 2 but
with responses reversed (i.e., they pressed the left key for
agentic words and the right key for communal words),
and (e) participants responded to female names and
agentic words with the left key and male names and com
munal words with the right key (abbreviated as female +
agency). The IAT effect is computed by subtracting the
mean male + agency response latency (Step 3) from the
mean female + agency response latency (Step 5). Thus,
positive difference scores reflect an automatic associa
tion between men and agency and women and
communality (i.e., implicit stereotypes). The order in
which participants performed Step 3 and Step 5 was
counterbalanced across participants.
Steps 3 and 5 of
the IAT are preceded by a practice block (n = 20 trials) so
that participants become familiar with the concepts
under investigation before proceeding to the experi
mental trials (n = 40 trials). The protocol for the gender
role and gender authority IATs paralleled the procedure
Self-report measures. Participants completed three
explicit measures of gender beliefs, designed to overlap
1318 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
with the IAT measures. These 7-point scales were
anchored by the endpoints of –3 (more true of women) and
3(more true of men). The gender roles index assessed the
concepts of career, occupation, homemaking, and child
care (α = .74). The gender authority index assessed the
concepts of authority, boss, leader, expert, subordinate,
and assistant (α = .69). The gender stereotype measure
(α = .78) assessed five agentic traits (individualistic, com
petitive, independent, hierarchical, and self-sufficient)
and five communal traits (communal, interdependent,
supportive, kinship-oriented, and connected). In each
case, mean judgments of feminine concepts were sub
tracted from mean judgments of masculine concepts.
Thus, high scores indicated more traditional gender
beliefs. These indexes had a possible range of –6 (nontra
ditional judgment) to 6 (traditional judgment).
Participants also completed the Ambivalent Sexism
Inventory (ASI) (Glick & Fiske, 1996) and a measure of
feminist identification (Henderson-King & Stewart,
1997). The ASI consists of two 11-item subscales that
assess hostile sexism (e.g., “Women seek to gain power by
getting control over men”; “Feminists are seeking for
women to have more power than men”) and benevolent
sexism (e.g., “Women should be cherished and pro-
tected by men”). Participants indicated agreement with
ASI items on a scale ranging from 0 (strongly disagree)to5
(strongly agree). High scores on ASI subscales reflect more
hostile or benevolent sexism. Because several items on
the hostile sexism scale pertain to gender and power, we
were principally interested in this subscale as a predictor
of attitudes toward female authority (Glick et al., 1997).
Participants also completed a social identity measure by
indicating on scales ranging from 1 (not at all)to5
(extremely) the extent to which they psychologically iden
tified with 17 social groups (e.g., students, liberals, con
servatives, fraternities, sororities, career women, and
feminists). The item of interest was participants’ femi
nist identification (Henderson-King & Stewart, 1997).
Because this item was related to identifying with career
women, r(67) = .62, p < .001, we combined these two
items to form a feminist identification index.
Volunteers were met by the experimenter and
escorted to a soundproof cubicle equipped with an
IBM-compatible PC. Participants first completed all
explicit measures in randomized order. They then com
pleted the priming measure and the IAT measures, in
counterbalanced order. In addition, the order in which
they completed the three IAT tasks was counterbalanced
(there were six possible orders). These counterbalanced
procedural variables did not significantly influence
for Implicit Measures
The data for each trial included response latency (in
ms) for correct responses and accuracy (i.e., whether
participants initially made an error). Data reduction fol
lowed prior procedures (Greenwald et al., 1998). To cor
rect for anticipatory responses and momentary inatten
tion, response latencies greater than 3,000 ms and less
than 300 ms were recoded as 3,000 and 300 ms, respec
tively. In addition, the first trial of each block was
dropped because of its typically lengthened latency.
Latencies were log-transformed to employ a statistic that
has satisfactory distribution of variance for analyses.
Analyses of participants’ accuracy revealed low error
rates on critical trials (Ms = 5% for IATs, 6% for the prim
ing measure). The priming measure excluded any trial
on which participants made an error (Fazio et al., 1995),
whereas error trials were included on IAT measures
(Greenwald et al., 1998).
Priming measure. First, baseline scores for each adjec-
tive were formed by averaging each participant’s Phase 1
latencies. Facilitation scores were then formed by sub-
tracting the mean latency for each primed target adjec-
tive from that adjective’s baseline score (i.e., baseline –
critical trial latency). These facilitation scores were
then averaged separately for positive and negative
adjectives within each prime type (high-authority
female, high-authority male, low-authority male, and
low-authority female). Figure 2 shows the results sepa
rately by participant gender. As can be seen, these facili
tation indexes showed positive scores (i.e., there was
facilitation for both positive and negative adjectives on
the primed trials, compared to baseline trials).
The facilitation scores were analyzed in a 2 (prime
gender) × 2 (prime authority) × 2 (valence) × 2 (partici
pant gender) mixed-model ANOVA with repeated mea
sures on all but the last factor. As suggested by Figure 2,
there was a main effect for prime gender such that facili
tation was greater for female than male primes, col
lapsed across valence. In addition, a Prime Gender ×
Valence interaction emerged, F(1, 67) = 4.11, p < .05.
Collapsed across authority, there was greater negative
than positive facilitation for female primes, t(67) = 2.89,
p < .01, whereas this difference was negligible for male
primes, t < 1.00. This finding shows a generalized preju
dice effect against female primes (i.e., implicit sexism).
Rudman, Kilianski / FEMALE AUTHORITY ATTITUDES 1319
The effect of interest, the three-way within-partici-
pants interaction, was robust, F(1, 67) = 15.31, p < .001. A
Prime Gender × Valence interaction was shown for
high-authority primes, F(1, 67) = 13.79, p < .001. Simple
effects revealed greater negative than positive facilita-
tion for female authorities, t(67) = 5.11, p < .001 (Ms=
295 vs. 141). By contrast, there was no reliable difference
between negative and positive facilitation scores for
male authorities, t(67) < 1.00 (Ms = 116 vs. 101). This
effect reflects more negative attitudes toward female
than male authorities (i.e., authority sexism). The Prime
Gender × Valence interaction was marginally reliable for
the low-status primes, F(1, 67) = 3.17, p < .09. In this case,
negative facilitation was greater than was positive facilita
tion for low-authority males (Ms = 176 vs. 112), whereas
facilitation scores were similar for low-authority females
(Ms = 234 vs. 227).
To provide a single-index attitude estimate, we com
puted the difference between the negative and positive
facilitation scores for each prime type (i.e., the white bar
latencies were subtracted from the black bar latencies
shown in Figure 2). Thus, high scores on this measure
reflect implicit prejudice (Wittenbrink, Judd, & Park,
1997). Table 1 shows the mean attitude effect for each
prime type. Also shown in Table 1 are the separate means
for men and women. As can be seen, women’s attitudes
toward female authorities did not differ significantly
from men’s. This finding is consistent with the literature
on implicit gender stereotypes (Greenwald & Banaji,
1995) and extends it to implicit attitudes toward female
authorities. By contrast, women showed more positive
attitudes toward low-authority primes than did men (see
Tests of whether each attitude estimate differed signif
icantly from zero were reliable for female authorities,
t(67) = 5.11, p < .001, and marginally reliable for
low-authority males, t(67) = 1.99, p = .05. In contrast,
these tests were nonsignificant for high-authority males
and low-authority females, both ts < 1.00. Thus, evidence
for implicit prejudice emerged most strongly for female
authorities. Finally, we compared attitudes toward
female authorities with attitudes toward each of the
other primes. Tests of each contrast were reliable, all
Fs(1, 67) > 4.65, ps < .05. In sum, the findings suggest that
attitudes toward female authorities were, on average,
negative for both men and women and more negative
than were attitudes toward the other primes.
IAT gender beliefs. Figure 3 displays the gender role,
gender authority, and gender stereotype IAT results sep
arately by participant gender. The IAT effect is repre-
sented by subtracting mean latencies for tasks compati-
ble with traditional beliefs (shown as white bars) from
mean latencies for tasks noncompatible with traditional
beliefs (shown as black bars) for each measure. Thus,
positive scores represent faster performance when per-
forming traditional, compared to nontraditional, tasks.
In general, Figure 3 shows this pattern. Tests of whether
these effects differed significantly from zero were reli-
able for each measure, all ts > 3.80, ps < .001.
Logged IAT effect scores were analyzed in a 3 (IAT
task) × 6 (IAT task order) × 2 (combined categorization
task order) × 2 (participant gender) mixed-model
ANOVA, with repeated measures on the first factor. A
main effect for IAT task was shown, F(2, 122) = 5.67, p <
.01. Figure 3 shows that the effect for the gender role IAT
(M = 169) was stronger than the effects for the gender
authority (M = 96) and gender stereotype (M = 100) IATs.
However, this finding was qualified by an IAT Task × Par
ticipant Gender interaction, F(2, 122) = 9.25, p < .001. As
shown in Table 1, men scored higher than women on the
gender authority IAT.
In contrast, there were no reliable
gender differences on the gender role and gender ste
reotype IATs (see Table 1). No other effects emerged in
this analysis. The effects for the counterbalanced proce
dural variables were nonsignificant, all Fs < 1.66, ps > .17.
GAM. Table 1 presents summary statistics for the
explicit attitude measure. High scores on this measure
reflect preference for male versus female authorities
(e.g., doctors, lawyers, professors). As expected, men
1320 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
Figure 2 Mean positive and negative facilitation scores (N = 69) as a
function of participant gender and prime type.
NOTE: Positive scores on the negative facilitation indexes represent fa
cilitation (compared to baseline latencies) when judging negative ad
jectives paired with representative primes. Positive scores on the
positive facilitation indexes represent facilitation (compared to base
line latencies) when judging positive adjectives paired with representa-
tive primes. Error bars are 95% confidence intervals for the
participants contributing to each mean (34 men, 35 women).
scored higher than women on this measure, t(67) = 5.00,
p < .001 (Ms = 3.27 vs. 2.61).
Gender beliefs. Table 1 presents data for the explicit
gender role, gender authority, and gender stereotype
measures. High scores on each measure correspond to
traditional gender beliefs. As can be seen, men scored
higher than women on the gender authority index,
t(67) = 2.11, p = .04. That is, men associated men with
high authority and women with low authority more than
did women. In contrast, no reliable gender differences
emerged on the gender role or gender stereotype rat
ings. These findings parallel those for the IAT measures.
Although the means for each measure are relatively low
(compared to an upper limit of 6.00), tests of whether
each mean differed significantly from the neutral point
were reliable, all ts > 9.00, ps < .001.
Additional measures. Table 1 also shows the results of
the ASI Hostile, ASI Benevolent, and feminist identifica
tion measures. In each case, gender differences were
found. Men were more likely to show evidence of hostile
sexism (e.g., antifeminism) and benevolent sexism (e.g.,
putting women on a pedestal) than were women. In con
trast, women were more likely to psychologically identify
with feminists and career women than were men.
Rudman, Kilianski / FEMALE AUTHORITY ATTITUDES 1321
TABLE 1: Summary Statistics for Implicit and Explicit Measures
Measure Mean Men’s Mean Women’s Mean Pooled SD t p
+154 (d = .64) +186 (d = .78) +121 (d = .50) 239.7 1.11 .27
+15 (d = .07) +22 (d = .08) +7 (d = .06) 262.6 .17 .87
+64 (d = .33) +144 (d = .59) –15 (d = –.06) 243.7 2.62 .01
+7 (d = .04) +86 (d = .33) –70 (d = –.26) 262.6 2.38 .02
Implicit Association Test (IAT) measures
+169 (d = .95) +134 (d = .75) +204 (d = 1.15) 177.7 1.64 .10
+96 (d = .53) +171 (d = .96) +20 (d = .11) 177.6 4.04 .001
+100 (d = .54) +85 (d = .46) +115 (d = .62) 184.0 .72 .47
Gender and Authority Measure (GAM)
2.94 3.27 2.61 .52 5.00 .001
2.32 2.52 2.11 1.27 1.32 .19
1.92 2.23 1.61 1.23 2.11 .04
1.06 .89 1.23 .97 1.49 .14
Ambivalent Sexism Inventory (ASI) hostile 2.37 2.90 1.84 .88 2.95 .004
ASI benevolent 2.74 3.01 2.47 .77 5.39 .001
Feminist identification 2.44 2.09 2.80 1.02 2.90 .005
NOTE: Priming and IAT scores are in ms. Gender differences were examined via t tests (df = 67). Effect sizes are Cohen’s d. Effect sizes were com
puted using the pooled standard deviation for men and women. Large, moderate, and small effect sizes correspond to ds of .8, .5, and .2, respectively
a. High scores indicate negative attitudes toward female authorities.
b. High scores indicate negative attitudes toward male authorities.
c. High scores indicate negative attitudes toward low-authority males.
d. High scores indicate negative attitudes toward low-authority females.
e. High scores indicate that, compared to women, men were more associated with careers.
f. High scores indicate that, compared to women, men were more associated with authority work roles.
g. High scores indicate that, compared to women, men were more associated with agentic traits.
Figure 3 Mean Implicit Association Test (IAT) latency results (N =
69) as a function of participant gender and IAT (gender
roles, gender authority, and gender stereotypes).
NOTE: Only latencies for the traditional and nontraditional tasks are
shown. Practice blocks and single categorization blocks are not in
cluded in the figure. Data are collapsed across counterbalanced proce
dural variables, which did not have significant influences on IAT
effects. Error bars are 95% confidence intervals for the participants
contributing to each mean (34 men, 35 women).
Relationships Among Explicit Measures
Table 2 presents the correlations among the explicit
measures. Because gender differences in these measures
could inflate these relations, we report first-order corre
lations, controlling for participant gender.
Explicit attitude relations. The top half of Table 2 shows
the relations among the explicit attitude measure
(GAM), hostile sexism, and benevolent sexism (ASI
subscales). As can be seen, preference for male authori
ties (GAM) was related to hostile sexism, r(66) = .38, p <
.001. Thus, antifemale attitudes covaried with prejudice
against female authorities (see also Glick et al., 1997). In
contrast, benevolent sexism was only weakly related to
preference for male authorities. Finally, the hostile and
benevolent subscales were related, consistent with Glick
and Fiske’s (1996) conceptualization of sexism as an
Explicit gender belief relations. Table 2 also shows the
relations among the explicit gender authority, gender
role, and gender stereotype measures. As can be seen,
gender authority and gender role beliefs were strongly
related, r(66) = .57, p < .001. This finding supports social
role theory’s argument that traditional labor divisions
(career vs. domestic roles) are linked to status and
authority differences for men and women (Eagly, 1987).
No other significant relations emerged. Although social
role theory posits that gender stereotypes stem from tra-
ditional labor divisions and that stereotypes, in turn, sup-
port role and authority differences, these relations were
not shown. Instead, gender stereotypes appeared to be
independent of gender role and authority beliefs, using
explicit measures (cf. Eagly & Steffen, 1984).
Explicit attitude-beliefs relations. The bottom half of
Table 2 shows the focal relations among prejudice
against female authorities (GAM) and explicit gender
beliefs. The primary aim of this analysis was to examine
support for the gender role, gender authority, and gen
der stereotype hypotheses. As can be seen, prejudice
against female authorities was related to gender author
ity beliefs, r(66) = .31, p < .01. Individuals who associated
men with high-authority roles (e.g., boss, leader, expert)
and women with low-authority roles (e.g., subordinate,
clerk, aid) also showed preference for male authorities
across a variety of domains (e.g., politics, education,
health). In contrast, the GAM was weakly related to gen
der role beliefs and negligibly related to gender stereo
types. In sum, the results support the gender authority
hypothesis, whereas no support was shown for the gen
der role and gender stereotype hypotheses, using
Feminist identification. The bottom row of Table 2
shows that feminist identification was related to the
GAM, r(66) = –.26, p < .05, and to ASI Hostile, r(66) =
–.25, p < .05. Thus, individuals who identified with femi
nists and career women were less likely to report preju
dice against female authorities and antifemale attitudes.
No other significant relations emerged.
Contrast scores. For the implicit analysis, we used four
single-index attitude estimates that reflected both our
conceptual definition of implicit prejudice and the reli-
able interactions shown by the facilitation score data
(Wittenbrink et al., 1997). The contrasts we computed,
and the effects they represent, are shown in Table 3. The
first contrast, female authority prejudice, is the differ
ence between facilitation for negative versus positive
adjectives when primes were female authorities. The sec
ond index, authority sexism, contrasts attitudes toward
female versus male authorities. It reflects the Prime
Gender × Valence interaction shown for high-authority
primes. The third contrast, relative female authority
prejudice, represents the Prime Gender × Prime Author
ity × Valence interaction. The advantage of this index is
that it provides a measure of female authority preju
dice relative to attitudes toward all other primes.
fourth and final contrast, implicit sexism, reflects the
Prime Gender × Valence interaction found in the three-
way ANOVA. It represents more negative attitudes
toward female primes compared to male primes. Table 4
presents the first-order correlations among the implicit
measures (controlling for participant gender).
Implicit attitude relations. Although each contrast has
been uniquely named, the implicit attitude indexes were
derived from the same facilitation score data. Therefore,
they were somewhat redundant, average r(67) = .36.
1322 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
TABLE 2: First-Order Correlations Among Explicit Measures
Gender Attitudes Gender Beliefs
Measure 1 2 3 4 5 6
Explicit gender attitudes
1. Gender and Authority
2. Ambivalent Sexism
Inventory (ASI) hostile .38**
3. ASI benevolent .18 .30*
Explicit gender beliefs
4. Gender authority beliefs .31** .19 .11
5. Gender role beliefs .12 .06 .09 .57**
6. Gender stereotypes –.08 .05 .03 .18 .06
7. Feminist identification –.26* –.25* –.14 .01 .06 .15
NOTE: Correlations are partial, controlling for participant gender
(df = 66).
*p < .05. **p < .01.
However, participant gender was unrelated to each mea
sure, all rs < .15, ns.
Implicit gender belief relations. The relations among the
gender authority, gender role, and gender stereotype
IATs were examined, controlling for participant gender.
As with the explicit measures, implicit gender authority
and gender role beliefs covaried, r(66) = .33, p < .01. This
result extends the conceptual link between gender roles
and gender authority beliefs to the implicit level (Eagly,
1987). No other significant relations emerged. As with
the explicit measures, gender stereotypes were indepen-
dent of gender role and gender authority beliefs (rs = .08
and .14, respectively).
Implicit attitude-belief relations. The top half of Table 4
shows the focal relations among the implicit attitude and
gender belief measures. The goal was to examine sup
port for the gender belief hypotheses using response
latency methods. As can be seen, the gender authority
IAT covaried positively with all four implicit attitude
indexes. By contrast, the relations between implicit atti
tudes and the gender role IAT were unreliably positive,
except in the case of implicit sexism. Furthermore, the
relations between implicit attitudes and the gender ste
reotype IAT were unreliably negative, except in the case
of implicit sexism.
Consistent with the explicit findings, these results
support the gender authority hypothesis. Whether atti
tudes toward female authorities were considered alone
or as a relative index, they were significantly related to
implicit gender authority beliefs. By contrast, the gender
role and gender stereotype hypotheses were not sup-
ported. Instead, implicit gender role beliefs showed
weakly positive links and gender stereotypes showed
weakly negative links to implicit attitudes toward female
However, all three gender belief IATs covaried with
implicit sexism. Individuals who automatically associ-
ated men (more than women) with careers and author
ity also showed more implicit dislike for women. By con
trast, individuals who automatically stereotyped women
(more than men) as communal showed less implicit sex
ism. These findings may reflect greater liking for women
on the part of individuals who automatically view them as
nice in an implicit version of the “women are wonderful”
effect (Eagly & Mladinic, 1993). Nonetheless, gender
stereotypes did not favorably influence implicit attitudes
toward female authorities, perhaps because nontradi
tional women are subtyped out of the superordinate,
protected category (Glick et al., 1997).
Implicit and Explicit Measures
The bottom half of Table 4 shows generally weak rela
tions among the implicit attitude and explicit measures,
as expected (Greenwald & Banaji, 1995). Specifically,
the first-order correlations between the GAM and the
four implicit attitude measures ranged from .09 to .15
(average r = .12). However, the female authority preju
dice index was reliably related to feminist identification,
r(66) = –.37, p < .01, and marginally related to hostile sex
Rudman, Kilianski / FEMALE AUTHORITY ATTITUDES 1323
TABLE 3: Contrast Weights Used in Priming Task Analysis
High Authority Low Authority
Female Male Female Male
Facilitation Score Contrast Prime Prime Prime Prime
1. Female authority prejudice
Positive adjectives –1 0 0 0
Negative adjectives 1 0 0 0
2. Authority sexism
Positive adjectives –1 1 0 0
Negative adjectives 1 –1 0 0
3. Relative female
Positive adjectives –1 1 1 –1
Negative adjectives 1 –1 –1 1
4. Implicit sexism
Positive adjectives –1 1 –1 1
Negative adjectives 1 –1 1 –1
a. Prime Gender × Valence interaction for high-authority primes.
b. Prime Gender × Prime Authority × Valence interaction.
c. Prime Gender × Valence interaction (collapsed across authority).
First-Order Correlations Among Implicit Attitude and
Authority Authority Authority Implicit
Measure Prejudice Sexism Prejudice Sexism
Implicit gender beliefs
Gender authority .38** .28* .27* .25*
Gender role .14 .15 .01 .26*
Gender stereotypes –.12 –.11 –.07 –.27*
Gender and Authority
Measure (GAM) .14 .15 .09 .11
Inventory (ASI) hostile .23 .18 .13 .21
ASI benevolent .12 –.08 .20 .31**
Gender authority .21 .18 .14 .19
Gender role .17 .03 .06 .07
Gender stereotypes .08 .06 .12 .18
Feminist identification –.37** –.15 –.05 –.17
NOTE: Correlations are partial, controlling for participant gender (df =
66). Correlations involving implicit measures were computed using
log-latency contrast scores. Correlations with untransformed latencies
*p < .05. **p < .01.
ism, r(66) = .23, p < .09. These results provide known
groups validity for this specific attitude estimate by show-
ing that hostile sexists and feminists were more and less
likely, respectively, to show implicit prejudice against
female authorities. The relations between feminist iden-
tification and ASI Hostile and the remaining attitude
indexes (i.e., authority prejudice, relative female author-
ity prejudice, and implicit sexism) were unreliable, all
ps > .12. Finally, the implicit sexism index was related to
ASI Benevolent, r(66) = .31, p < .01, suggesting that indi-
viduals who endorsed paternalistic attitudes toward
women also implicitly disliked women.
Table 5 shows the relationships among the IAT and
explicit measures. As can be seen, these relations were
nonsignificant, with two exceptions. ASI Benevolent was
related to the gender status and gender stereotype IATs,
suggesting that benevolent sexists were likely to automat
ically associate men with high-authority and women with
low-authority roles and to implicitly stereotype men as
agentic and women as communal.
of Implicit Measures
Because response latency and self-report measures of
attitudes and beliefs typically show weak relations
(Greenwald & Banaji, 1995), the construct validity of
implicit measures may be best assessed using different
response latency methods. To date, this strategy has not
been tested. The present research uniquely examined
the hypothesis that implicit measures assess similar phe
nomena (i.e., automatic associations in long-term mem
ory). The priming task assessed evaluative associations
(i.e., attitudes), whereas the IAT assessed semantic asso
ciations (i.e., beliefs). Nonetheless, the conceptual link
between attitudes and beliefs was expected to emerge,
and it did, as each implicit attitude index covaried with
IAT-assessed gender authority beliefs. Furthermore, the
implicit sexism index covaried with each IAT, showing
that gender authority, role, and trait associations are
each important correlates of generalized, implicit preju
dice. The positive relations among implicit sexism, gen
der role, and gender authority beliefs are consistent with
social role theory (Eagly, 1987), because traditional
labor divisions are thought to underlie status and prestige
differences that evaluatively favor men. The negative rela
tionship between implicit stereotypes and sexism is con
sistent with research showing that agency-communality
beliefs enhance women’s overall likability (Eagly &
Mladinic, 1993). Taken together, these findings extend
basic tenets of social role theory and recent gender atti
tude research to the implicit level and, in so doing, sup
port the construct validity of both the priming and IAT
Construct Validity of
Explicit Attitude Measure
The known groups validity of the GAM was shown by
its relations with participant gender, feminist identifica-
tion, and hostile sexism. Women and individuals who
identified with feminists and career women showed less
prejudice against female authorities, whereas hostile
sexists showed more (Forsythe et al., 1997; Glick et al.,
1997). The GAM also was related to explicit gender
authority beliefs but not to explicit gender role beliefs or
gender stereotypes. This pattern supports the specificity
of implicit attitude-belief relations. It is not the case that
relations between female authority attitudes and gender
role or trait beliefs were undiscoverable using automatic
methods. Rather, it appears that these relations did not
exist at either the implicit or explicit level. In contrast,
specific beliefs linking men to high authority and
women to low authority were related to prejudice against
female authorities, using both implicit and explicit mea
sures (Fishbein & Ajzen, 1975).
To date, response-latency methods have been applied
to gender stereotypes (e.g., Banaji & Hardin, 1996;
Blair & Banaji, 1996; Rudman et al., 1999) but not to
gender attitudes. Attitudes toward female authority were
the focus of the present research because of their impli
cations for gender discrimination in the workplace (i.e.,
preservation of the economic status quo). The assess
ment strategy included both implicit and explicit atti
tudes, which showed the predicted dissociation. This dis
sociation may be due to cognitive (i.e., an inability to
access information) (Greenwald & Banaji, 1995) or stra
tegic (i.e., impression management) (Dovidio & Fazio,
1324 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
TABLE 5: First-Order Correlations Among Implicit Association
Test (IAT) and Explicit Measures
Gender Gender Gender
Explicit Measure Authority Role Stereotype
Gender authority .16 .12 .02
Gender role .02 .09 .20
Gender stereotypes .22 –.11 .16
Gender and Authority
Measure (GAM) .21 .05 .18
Inventory (ASI) hostile .13 .05 .05
ASI benevolent .30* .08 .25*
Feminist identification –.05 .06 –.09
NOTE: Correlations are partial, controlling for participant gender
(df = 66). Correlations involving implicit measures were computed us
ing log-latency difference scores. Correlations with untransformed la
tencies were similar.
*p < .05. **p < .01.
1992) explanations. In either case, they underscore the
need for both automatic and controlled measures to
explore the nuances of attitudes toward female
in Female Authority Attitudes
Consistent with the literature on implicit gender ste
reotypes, implicit attitudes toward female authorities
were similar for men and women (Banaji & Hardin,
1996; Blair & Banaji, 1996; Greenwald & Banaji, 1995).
Defining implicit attitudes as the specific contrast
between negative and positive facilitation for female
authority primes, we found that attitudes toward female
authorities were negative for men and women alike and
more negative than were attitudes toward the other
primes (male authorities, low-authority females, and
low-authority males). In contrast, gender differences
were shown on the explicit attitude measure, with
women reporting less prejudice against female authori
ties than men. This finding supports a large literature
showing that women’s explicit attitudes are more egalitar-
ian than are men’s (e.g., Glick & Fiske, 1996; Swim et al.,
1995; Williams & Best, 1990).
As individual difference measures, feminist identifi-
cation and hostile sexism performed more consistently.
Self-reported feminists showed less explicit prejudice
against female authorities (Forsythe et al., 1997) and
also less implicit prejudice on the female authority preju-
dice measure. In contrast, hostile sexists showed more
explicit prejudice against female authorities (Glick et al.,
1997) and (marginally) more implicit prejudice on the
female authority prejudice measure. These findings sug
gest that feminist identification and antifemale attitudes
can have positive and negative effects, respectively, on
accepting female authority, both consciously and
Gender Beliefs and Female
The primary aim of the research was to discern why
men are more readily accepted than women in positions
of power. In Eagly et al.’s (1992) meta-analysis of leader
evaluations, the authors concluded that female leaders
“ ‘pay a price’ in terms of relatively negative evaluations
if they intrude on traditionally male domains by...occu
pying male-dominated leadership positions” (p. 18).
This finding has serious implications because, beyond
leading the local chapter of the Daughters of the Ameri
can Revolution, female authorities are ipso facto intrud
ing on male territory, historically speaking. Therefore, it
is important to uncover correlates of negative reactions
to female authority.
Three potential correlates were examined: gender
role associations (i.e., beliefs that men and women’s
social roles differ), gender authority associations (i.e.,
beliefs that men have more authority than women), and
gender stereotypes (i.e., beliefs that men are more
agentic than women). Results showed consistent sup
port for the gender authority hypothesis. Associating
men with high authority and women with low authority
covaried with negative attitudes toward female authority.
This was true whether attitudes were measured implicitly
or explicitly. Although, on average, women showed less
susceptibility to gender authority associations than did
men, the relationship between gender authority beliefs
and female authority prejudice was similar for both gen
ders, at both the implicit and explicit level.
These findings suggest that negative reactions to
female authority may stem, in part, from an implicit pro
totype for male leaders and the attendant belief that it is
more natural for men to take control. Individuals may be
comforted by male leadership for the simple fact that
they are accustomed to viewing men as authority figures
and women as subordinates. Thus, women who occupy
male-dominated leadership positions may be dis-
liked—both implicitly and explicitly—because they
breach expectancies that men (not women) occupy pow-
erful roles. However, the association between male gen-
der and leadership also is likely to be prescriptive,
because a tradition of male authority results in norms of
male entitlement (Jost & Banaji, 1994). Thus, female
leaders may be disliked for “stepping on men’s toes”
(i.e., usurping men’s position in the hierarchy). In sum,
female authorities may be disliked because they are unfa-
miliar and/or because they are viewed as threatening
intruders (Eagly et al., 1992). These interpretations are
distinct but undoubtedly related, because the descrip
tive (what is) and prescriptive (what should be) elements
of gender beliefs are intertwined (Glick & Fiske, 1999).
Because gender authority beliefs consistently
covaried with attitudes, it is important to note that these
beliefs, as assessed by the IAT and self-reports, were unre
liably related, r(66) = .16, p = .25 (see Table 5). This disso
ciation again bolsters the conceptual distinction
between automatic and conscious beliefs. However, it
bodes ill for gender equality because it shows that even
individuals who possess consciously egalitarian beliefs
are susceptible to implicit gender authority beliefs and,
hence, implicitly negative female authority attitudes.
In contrast, there was no support for the gender role
or gender stereotype hypotheses. That is, gendered asso
ciations regarding (a) career versus domestic roles and
(b) agentic versus communal traits were unreliably
related to implicit and explicit female authority atti
tudes. However, gender role and gender authority
beliefs covaried, using both implicit and explicit mea
Rudman, Kilianski / FEMALE AUTHORITY ATTITUDES 1325
sures, consistent with arguments that the traditional divi
sion of labor (career vs. domestic) connotes high- versus
low-status expectancies for men and women, respec
tively (Eagly, 1987).
Surprisingly, gender role and authority beliefs were
negligibly related to gender stereotypes, for both
implicit and explicit measures (cf. Eagly & Steffen,
1984). This finding may reflect the extent to which the
stereotype has taken on a life of its own (i.e., is no longer
dependent on labor divisions). However, this does not
explain why the stereotype did not predict attitudes
toward female authorities. One possibility is that the
agency-communality distinction may be less important
than other gender stereotypes, depending on the type of
female authority being evaluated. For example, reluc
tance to accept female expertise (e.g., scientists) may be
due more to beliefs about men’s superior intellect than
to their agency. Likewise, legitimate power differences
(e.g., policing and judgeships) may be buttressed by
beliefs about men’s superior strength or reasoning abili
ties. Future research should examine a broader range of
gender stereotypes to assess this possibility. Nonetheless,
if authority roles are construed as requiring agency, it
was not unreasonable to expect that people who view
men as agentic and women as communal might react
negatively to female authorities. Instead, people who
showed differential gender authority associations
reacted negatively to female authorities.
Individuals who associate men with high authority
and women with low authority are likely to show preju
dice against female authorities. Support for this relation
ship was found using both implicit and explicit mea
sures. The intermethod independence of these
measures shows that gender equality may be hindered by
both automatic and conscious gender authority associa
tions. That is, the implicit prototype for male authority is
also, for some individuals, an explicit prototype, and
both carry separate—but equally negative—implica
tions for female authorities. Although tested at both the
implicit and explicit level, gender role beliefs and gen
der stereotypes were unrelated to female authority atti
tudes. Thus, prejudice against female authority may be
due more to associations linking men to power and influ
ence than to role or trait expectancies. In other words,
women may be viewed as legitimate careerists, possessed
of the agency necessary for flying 747s and performing
surgery. However, if they violate expectancies that men
(not women) occupy powerful roles, their authority in
the cockpit or the operating room may not be welcomed.
The Gender and Authority Measure
1. If I were in serious legal trouble, I would prefer a male to a
2. The people I look up to most are women.
3. I would feel more comfortable if the pilot of an airplane
I was traveling on were male.
4. I would rather be stopped by a woman police officer (vs. a
5. I probably prefer that the U.S. president is a man, versus a
6. In general, I would rather work for a man than for a woman.
7. If I were having a serious operation, I would have more
confidence in a male surgeon.
8. When it comes to politics, I would rather vote for women
than for men.
9. For most college courses, I prefer a male professor to a
10. Personally, I would rather go to a male doctor than a female
11. In general, women make better leaders than men do.
12. In most areas, I would rather take advice from a man than
from a woman.
13. In general, I would rather take orders from a man than
from a woman.
14. If I were being sentenced in court, I would prefer that the
judge be a woman.
15. In general, I feel more comfortable when a man
(vs. a woman) is in charge.
a. Items require reverse scoring.
1. The adjectives’ valence scores were derived from Williams and
Best’s (1990) favorability norms. On this scale, scores greater than 500
reflect positive valence, whereas scores less than 500 reflect negative
valence. The adjectives’ gender association scores were derived from
Williams and Best’s (1990) Sex Stereotype Index. On this scale, scores
greater than 500 reflect association with male gender, whereas scores
less than 500 reflect association with female gender.
2. These pictures consisted of men and women shown in both tradi
tional and nontraditional roles. Pictures of women included an aero
bics instructor, housewife, telephone operator, architect, construction
worker, and paramedic. Pictures of men included a businessman, bar
tender, college student, telephone operator, model, and typist.
3. The scale was first administered to a sample of 191 respondents
(84 men, 107 women). Respondents also completed traditional and
contemporary measures of sexism (Glick & Fiske, 1996; Spence &
Helmreich, 1972) and a measure of social desirability (Paulhus, 1984).
Exploratory factor analysis of the Gender and Authority Measure
(GAM) suggested a single factor structure, and the internal consis
tency for the combined scale was adequate (α = .88). The pilot study
revealed significantly higher scores for men than women and reliably
positive relationships with the sexism measures, average r = .32, all ps<
.001. Thus, the scale showed evidence of both known groups and con
vergent validity. Its relationship to both the impression management
and self-deception subscales of the Balanced Inventory of Desirable
Responding (BIDR) (Paulhus, 1984) was nonsignificant, suggesting
1326 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN
4. We thank Tony Greenwald and Scott Tiernan for providing us
with stimuli for this Implicit Association Test (IAT).
5. Pilot respondents (N = 32) rated a list of 20 traits (10 agentic, 10
communal) on three semantic differential dimensions of good-bad,
harmful-harmless, and positive-negative. In addition, they rated the
extent to which each trait was more associated with men versus women.
The agentic and communal traits used in the gender stereotype IAT
(and explicit gender stereotype measure, described below) were
selected on the basis of this pretest to be similar in valence but different
in stereotypic association.
6. Nonorthogonally, key assignment for Step 2 also was counterbal
anced. For example, participants who performed the female + agency
task first also pressed the left key for agentic words and the right key for
communal words in Step 2.
7. An unexpected Prime Authority × Valence × Participant Gender
interaction also emerged, F(1, 67) = 4.73, p < .05. The Prime Authority ×
Valence interaction was reliable for women (p = .01). Specifically,
women showed more positive facilitation for low-authority primes
compared to high-authority primes, t(67) = 3.13, p < .01, whereas their
negative facilitation scores for these primes were similar (t < 1.00). By
contrast, analysis of the men’s data showed only a main effect for
valence, such that men’s negative facilitation scores were greater than
their positive facilitation scores (p < .001). These findings suggest a
positivity bias toward low-status targets on the part of women and a gen
eralized negativity bias on the part of men. Because these unexpected
effects are peripheral to the present article’s focus, we will not discuss
them. Participant gender did not reliably interact with other variables,
all ps > .09.
8. This finding is consistent with past research showing that the
evaluative connotations of judgments influence sex differences in
implicit gender stereotypes (Rudman et al., 1999). Because
high-authority roles are more valued than low-authority roles, women’s
implicit self-esteem may inhibit them from making judgments that
connote inferiority for their gender (see also Greenwald et al., in
9. We thank an anonymous reviewer for this suggestion.
Ashmore, R. D., Del Boca, F. K., & Bilder, S. M. (1995). Construction
and validation of the Gender Attitude Inventory, a structured
inventory to assess multiple dimensions of gender attitudes. Sex
Roles, 32, 753-785.
Banaji, M., & Greenwald, A. G. (1995). Implicit gender stereotyping in
judgments of fame. Journal of Personality and Social Psychology, 68,
Banaji, M. R., & Hardin, C. D. (1996). Automatic stereotyping. Psycho
logical Science, 7(3), 136-141.
Berger, J., Fisek, M. H., Norman, R. Z., & Zelditch, M., Jr. (1977). Status
characteristics and social interaction: An expectations states approach.
New York: Elsevier Science.
Blair, I. V., & Banaji, M. R. (1996). Automatic and controlled processes
in stereotype priming. Journal of Personality and Social Psychology, 70,
Breckler, S. J. (1984). Empirical validation of affect, behavior, and
beliefs as distinct components of attitude. Journal of Personality and
Social Psychology, 47, 1191-1205.
Carli, L., & Eagly, A. H. (in press). Gender effects on social influence
and emergent leadership. To appear in G. N. Powell (Ed.), Hand
book of gender in organizations. Thousand Oaks, CA: Sage.
Cohen, J. (1988). Statistical power for the behavioral sciences. Hillsdale, NJ:
Dovidio, J. F., & Fazio, R. H. (1992). New technologies for the direct and
indirect assessment of attitudes. In J. M. Tanur (Ed.), Questions
about questions: Inquiries into the cognitive bases of surveys (pp. 204-
237). New York: Russell Sage.
Eagly, A. H. (1987). Sex differences in social behavior: A social role interpreta
tion. Hillsdale, NJ: Lawrence Erlbaum.
Eagly, A. H., & Karau, S. J. (1991). Gender and the emergence of lead
ers: A meta-analysis. Journal of Personality and Social Psychology, 60,
Eagly, A. H., Makhijani, M. G., & Klonsky, B. G. (1992). Gender and the
evaluation of leaders: A meta-analysis. Psychological Bulletin, 111,
Eagly, A. H., & Mladinic, A. (1993). Are people prejudiced against
women? Some answers from research on attitudes, gender stereo
types, and judgments of competence. In W. Stroebe & M. Hewstone
(Eds.), European review of social psychology. New York: John Wiley.
Eagly, A. H., & Steffen, V. J. (1984). Gender stereotypes stem from the
distribution of men and women into social roles. Journal of Personal
ity and Social Psychology, 46, 735-754.
Eagly, A. H., & Wood, W. (1991). Explaining sex differences in social
behavior: A meta-analytic perspective. Personality and Social Psychol
ogy Bulletin, 17, 306-315.
Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995). Vari
ability in automatic activation as an unobtrusive measure of racial
attitudes. A bona fide pipeline? Journal of Personality and Social Psy
chology, 69, 1013-1027.
Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention, and behavior:
An introduction to theory and research. Reading, MA: Addison-Wesley.
Forsythe, D. R., Heiney, M. M., & Wright, S. S. (1997). Biases in apprais
als of women leaders. Group Dynamics, 1, 98-101.
French, J., Jr., & Raven, B. (1959). The bases of social power. In D. Cart
wright (Ed.), Studies in social power (pp. 150-167). Ann Arbor: Uni
versity of Michigan.
Gallup, G. H. (1996). The Gallup poll. Wilmington, DE: Scholarly
Glick, P., Diebold, J., Bailey-Werner, B., & Zhu, L. (1997). The two faces
of Adam: Ambivalent sexism and polarized attitudes toward
women. Personality and Social Psychology Bulletin, 23, 1323-1334.
Glick, P., & Fiske, S. T. (1996). The Ambivalent Sexism Inventory: Dif-
ferentiating hostile and benevolent sexism. Journal of Personality and
Social Psychology, 70, 491-512.
Glick, P., & Fiske, S. T. (1999). Sexism and other “isms”: Interdepen-
dence, status, and the ambivalent content of stereotypes. In W. B.
Swann, J. H. Langlois, & L. A. Gilbert (Eds.), Sexism and stereotypes in
modern society: The gender science of Janet Taylor Spence (pp. 193-221).
Washington, DC: American Psychological Association.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition:
Attitudes, self-esteem, and stereotypes. Psychological Review, 102,
Greenwald, A. G., Banaji, M. R., Rudman, L. A., Farnham, S. D., Nosek,
B. A., & Rosier, M. (in press). Prologue to a unified theory of atti
tudes, stereotypes, and self-concept. To appear in J. P. Forgas (Ed.),
Feeling and thinking: The role of affect in social cognition and behavior.
New York: Cambridge University Press.
Greenwald, A. G., McGhee, D., & Schwartz, J.L.K. (1998). Measuring
individual differences in implicit cognition: The Implicit Associa
tion Test. Journal of Personality and Social Psychology, 74, 1464-1480.
Heilman, M. E. (1983). Sex bias in work settings: The lack of fit model.
Research in Organizational Behavior, 5, 269-298.
Heilman, M. E., Block, C. J., Martell, R. F., & Simon, M. C. (1989). Has
anything changed? Current characterizations of men, women, and
managers. Journal of Applied Psychology, 74, 935-942.
Henderson-King, D., & Stewart, A. J. (1997). Feminist consciousness:
Perspectives on women’s experience. Personality and Social Psychol
ogy Bulletin, 23, 415-426.
Johnson, P. (1976). Women and power: Toward a theory of effective
ness. Journal of Social Issues, 32, 99-110.
Jost, J. T., & Banaji, M. R. (1994). The role of stereotyping in system-jus
tification and the production of false-consciousness. British Journal
of Social Psychology, 33, 1-27.
Lyness, K., & Thompson, D. (1997). Above the glass ceiling? A compar
ison of matched samples of female and male executives. Journal of
Applied Psychology, 82, 359-375.
Moore, G., & Shackman, G. (1996). Gender and authority: A
cross-national study. Social Science Quarterly, 77, 274-288.
Rudman, Kilianski / FEMALE AUTHORITY ATTITUDES 1327
Paulhus, D. L. (1984). Two-component models of socially desirable
responding. Journal of Personality and Social Psychology, 46, 598-609.
Reskin, B. F., & Ross, C. E. (1992). Jobs, authority, and earnings among
managers. Work and Occupations, 19, 342-65.
Rubner, M. B. (1991). More workers prefer a man in charge. American
Demographics, 13, 11.
Rudman, L. A. (1997). Gender and authority attitudes. Unpublished
Rudman, L. A., Greenwald, A. G., & McGhee, D. E. (1999). Self-esteem
and gender identity are manifest in implicit gender stereotypes. Manuscript
submitted for publication.
Rudman, L. A., Greenwald, A. G., Mellott, D. S., & Schwartz, J.L.K. (in
press). Measuring the automatic components of prejudice: Flexi
bility and generality of the Implicit Association Test. Social
Sonnert, G., & Holton, G. (1996). Career patterns of women and men
in the sciences. American Scientist, 84, 63-71.
Spence, J. T., & Helmreich, R. (1972). The Attitudes Toward Women
Scale: An objective instrument to measure attitudes toward the
rights and roles of women in contemporary society. JSAS Catalog of
Selected Documents in Psychology, 2, 66. (a)
Swim, J. K., Aikin, K. J., Hall, W. S., & Hunter, B. A. (1995). Sexism and
racism: Old-fashioned and modern prejudice. Journal of Personality
and Social Psychology, 68, 199-214.
Vianello, M., & Siemienska, R. (1990). Gender inequality: A comparative
study of discrimination and participation. Newbury Park, CA: Sage.
Watson, C. (1988). When a woman is the boss: Dilemmas in taking
charge. Group and Organization Studies, 13, 163-181.
Williams, J. E., & Best, D. L. (1990). Measuring sex stereotypes: A
multination study. Newbury Park, CA: Sage.
Wittenbrink, B., Judd, C. M., & Park, B. (1997). Evidence for racial prej
udice at the implicit level and its relationship with questionnaire
measures. Journal of Personality and Social Psychology, 72, 262-274.
Wright, F. (1976). The effects of style and sex of consultants and sex of
members in self-study groups. Small Group Behavior, 7, 433-456.
Wright, E. O., & Baxter, J. (1995). The gender gap in workplace author
ity: A cross-national study. American Sociological Review, 60, 407-435.
Received October 19, 1998
Revision accepted June 10, 1999
1328 PERSONALITY AND SOCIAL PSYCHOLOGY BULLETIN