Content uploaded by Tyler Okimoto
Author content
All content in this area was uploaded by Tyler Okimoto on Mar 22, 2016
Content may be subject to copyright.
Why Are Women Penalized for Success at Male Tasks?:
The Implied Communality Deficit
Madeline E. Heilman
New York University
Tyler G. Okimoto
Flinders University of South Australia
In 3 experimental studies, the authors tested the idea that penalties women incur for success in
traditionally male areas arise from a perceived deficit in nurturing and socially sensitive communal
attributes that is implied by their success. The authors therefore expected that providing information of
communality would prevent these penalties. Results indicated that the negativity directed at successful
female managers—in ratings of likability, interpersonal hostility, and boss desirability—was mitigated
when there was indication that they were communal. This ameliorative effect occurred only when the
information was clearly indicative of communal attributes (Study 1) and when it could be unambiguously
attributed to the female manager (Study 2); furthermore, these penalties were averted when communality
was conveyed by role information (motherhood status) or by behavior (Study 3). These findings support
the idea that penalties for women’s success in male domains result from the perceived violation of
gender-stereotypic prescriptions.
Keywords: prescriptive gender stereotypes, sex discrimination, penalties for success, communal prescrip-
tions
Research on sex bias in the workplace has demonstrated repeat-
edly that in situations in which performance level is ambiguous or
lends itself to distortion, women are not perceived to be as com-
petent as men in performing male gender-typed work (Dipboye,
1985; Heilman, 1983, 1995; Heilman, Martell, & Simon, 1988;
Swim, Borgida, Maruyama, & Myers, 1989). Consequently, am-
biguity about performance quality has been shown to result in
discrimination in selection, placement, and performance evalua-
tion (Cejka & Eagly, 1999; Heilman, 1995, 2001; Nieva & Gutek,
1980; Tosi & Einbinder, 1985). Recent evidence suggests, how-
ever, that even when unequivocal evidence exists that a woman is
successful in male gender-typed work, she faces career-hindering
problems in work settings—problems of being disliked and inter-
personally derogated (Heilman, Wallen, Fuchs, & Tamkins, 2004).
The research reported in this article is designed to provide further
insight into why these responses to successful women arise and the
conditions that can inhibit their occurrence. In particular, we
propose that negative reactions to women who are successful in
traditionally male domains are a consequence of the perception
that these women have violated stereotype-based “oughts” about
how women should behave and the resulting assumption that they
are deficient in feminine attributes.
In contrast to the descriptive component of gender stereotypes,
which designates what men and women are like, there is a pre-
scriptive component that designates what men and women should
be like (Burgess & Borgida, 1999; Eagly & Karau, 2002; Heilman,
2001; Rudman & Glick, 2001). The dictates of prescriptive sex
stereotypes are highly specific and widely shared. They specify
that women should behave communally, exhibiting nurturing and
socially sensitive attributes that demonstrate concern for others,
such as being kind, sympathetic, and understanding. They also
specify what women should not do— engage in behaviors typically
prescribed for men that are thought to be incompatible with the
behaviors prescribed for women. Thus, agentic behavior, behavior
that demonstrates dominance, competitiveness, and achievement
orientation, is generally considered out of bounds for women.
These gender-stereotyped prescriptions can affect reactions to
women. Because the behavioral prescriptions inherent in gender
stereotypes function as norms to be fulfilled, their perceived vio-
lation is likely to arouse disapproval and promote negativity, as do
counternormative behaviors more generally (Cialdini & Trost,
1998). Penalties are therefore likely to ensue.
Research has demonstrated that penalties indeed result when
women engage in behaviors that are counter to female stereotypic
prescriptions. These penalties have been found to take the form of
social rejection and personally directed negativity and to have
detrimental consequences for career-relevant organizational re-
wards. For example, Rudman (1998; Rudman & Glick, 2001)
found that women who engaged in self-promoting behavior were
judged as lacking social skills and were less likely to be recom-
mended for hiring. Results of other studies also have indicated that
when women behave in ways that are typically reserved for men in
our culture, they are less liked and found less socially acceptable
than men who behave in a similar manner or than women who
behave in more stereotype-consistent ways (Bartol & Butterfield,
1976; Carli, 1990; Carli, LaFleur, & Loeber, 1995; Jago & Vroom,
1982).
However, it is not necessary for women to actually behave
counternormatively to induce social penalties; the mere knowledge
that a woman has been successful in a male domain produces
Madeline E. Heilman, Department of Psychology, New York Univer-
sity; Tyler G. Okimoto, School of Psychology, Flinders University of
South Australia, Adelaide, Australia.
Correspondence concerning this article should be addressed to Madeline E.
Heilman, Department of Psychology, New York University, 6 Washington
Place, Room 576, New York, NY 10003. E-mail: madeline.heilman@nyu.edu
Journal of Applied Psychology Copyright 2007 by the American Psychological Association
2007, Vol. 92, No. 1, 81–92 0021-9010/07/$12.00 DOI: 10.1037/0021-9010.92.1.81
81
inferences that she has engaged in stereotype-violating behavior,
resulting in social penalties. Much of the research supporting this
idea has concerned women who are successful in the organiza-
tional role of manager, a role that is typically seen as male gender
typed (Powell, Butterfield, & Parent, 2002; Schein, 2001). Several
investigations have found that when research participants were
told only that female managers had been successful (with no
additional behavioral information supplied), they characterized
these managers as lacking the prescribed favorable interpersonal
qualities related to communality and as instead possessing traits
such as selfishness, deceitfulness, deviousness, coldness, and ma-
nipulativeness (Heilman, Block, & Martell, 1995; Heilman, Block,
Martell, & Simon, 1989; Heilman et al., 2004). There also has been
indication that female managers who are reported to be successful
are decidedly less liked than their male counterparts and less
preferred as bosses (Heilman et al., 2004). It thus appears to take
little more than the knowledge that a woman is successful at male
sex-typed work to instigate interpersonally negative reactions to
her. Moreover, the fact that this negativity directed at female
managers who are successful has not been found to occur when
their success is in an arena that is designated as female or neutrally
gender typed (Heilman et al., 2004, Study 2) lends support to the
idea that it is not success but rather the inferences success produces
about the violation of gender-stereotypic prescriptions that fuels
these negative reactions.
There is, however, lack of clarity about what, exactly, is at the
root of the penalties that women incur for their success in male
domains. Violating gender-stereotyped prescriptions by being suc-
cessful in these arenas has two different consequences. Most
directly, because agentic behavior is believed to be essential for
excellent performance of male gender-typed tasks regardless of the
sex of the performer, it indicates that a woman has done what men,
not women, are supposed to do— be agentic. However, because
communality and agency tend to be conceived of as oppositional,
agentic behavior also implies that a woman has failed to do what
women are supposed to do— be communal. Which of these gender
norm violations is the source of the disapproval and social rejec-
tion? We believe successful women’s assumed desertion of fem-
inine imperatives is critical for explaining the penalties they incur
and for understanding whether and how such penalties can be
averted.
Our reasoning is rooted in the nature of the penalties themselves.
Negative reactions to successful women tend to target interper-
sonal traits in the communal domain: These women are character-
ized as the antithesis of the female nurturer—as the quintessential
“bitch” who is concerned not at all about others but only about
herself (e.g., Heilman et al., 2004). These findings suggest that it
is the perceived deficiency in communality implied by a woman’s
success in a male job, not the perception of inappropriate agenti-
cism, that is the true irritant and the primary source of the disap-
proval driving the resulting social penalties. Thus, the woman’s
success, which in this situation demonstrates agenticism, may not
be the problem; rather, the trouble may be the implications of the
success for perceptions of her communality. If this is the case,
when the assumption of the communality violation is precluded, a
woman’s success in a nontraditional role should not produce
negative consequences.
The following studies test this idea. We propose that penalties
for a woman’s success in a male domain arise from the perceived
violation of communality prescriptions, and we therefore expect
that providing verification of communality will prevent the social
disapproval and social penalties otherwise directed at a woman
when she is successful in a male job. Because there is no commu-
nal prescription for men that is perceived to be violated when they
are successful in male jobs, we do not expect information about
communality to affect men’s evaluations. The first study directly
tests this proposition. In this study, participants reviewed informa-
tion about a woman and a man who were depicted as highly
successful at a male gender-typed managerial job, and additional
information about their communality was or was not communi-
cated. To demonstrate that communality information in particular,
not favorable information in general, is responsible for the averting
of penalties, we also included a control condition in which addi-
tional information was provided about the manager that was fa-
vorable but not communal in its content. We expect the following:
Hypothesis: Female managers who are successful in male
gender-typed jobs will be more disliked, elicit more negative
interpersonal characterizations, and be less preferred as
bosses than similarly successful male managers unless infor-
mation about their communality is provided.
Study 1
Method
Participants and Design
Participants were 75 male and female undergraduates recruited from an
introductory psychology course who participated in exchange for partial
fulfillment of course experiment participation requirements. The design
wasa2⫻ 3 mixed factorial with sex of the stimulus person (male or
female) a within-subject factor and type of information provided (no added
information, communal information, or positive noncommunal informa-
tion) a between-subjects factor. Participants were randomly assigned to one
of the three information conditions, and each reviewed both a male and a
female target.
Procedure
The experimenter informed the participants that the study concerned
relationships at work and, in particular, first impressions. They were told
that they would be asked to read about and evaluate 3 managers, randomly
selected from among a group of 10, who had recently been hired for a
similar position in different divisions of a large organization. Participants
were then given packets containing information about the position and
about each of the 3 managers and were asked to give their reactions to them
on a brief questionnaire. Two of the managers, 1 male and 1 female, were
our stimulus targets. After the questionnaires were completed and col-
lected, the experimenter gave both a written and a verbal debriefing that
revealed the purpose of the experiment and the manipulations used in the
study.
Stimulus Materials
The stimulus packet began with a checklist containing the names of 10
managers (9 male and 1 female), all ostensibly potential evaluatees, 3 of
whom had been designated as the managers to be reviewed by the partic-
ipant. The checklist was overwhelmingly male to reinforce the male-
dominated nature of the position. This was followed by a job description
summary indicating the position held by all 3 managers, vice president
(VP) of financial affairs, and the responsibilities of the job (e.g., supervis-
82
HEILMAN AND OKIMOTO
ing the financial affairs department, overseeing allocation of company
funds, directing financial affairs employees, preparing budget reports, and
leading work projects). The male gender-typed nature of the job was
communicated in the job summary via the type of job (top manager of a
finance department), demographics (the summary indicated that 88% of
these managers company-wide were male), and the types of job responsi-
bilities involved (e.g., managing, supervising, leading, directing,
overseeing).
The participants then read memos from the chief executive officer
(CEO) to the company employees introducing each VP as a new member
of his or her respective management group. The memos were described as
instrumental in the company’s efforts to introduce employees to incoming
upper level managers. Each memo was four paragraphs long and contained
background information regarding the new VP. Memos about the two
stimulus targets reported their educational background (undergraduate de-
gree from the University of Southern California or University of Michigan;
master’s in business administration from Northwestern University or the
University of California, Berkeley), city of origin (Cleveland or Saint
Louis), interests outside of work (movies, jogging, and photography or
reading, tennis, and travel), and past work experience (approximately 3
years in a financial manager position). The two stimulus targets were
depicted as having been highly successful at the last company they had
worked for, and they were described as having been awarded organiza-
tional recognition for their performance and commitment. Additionally,
there were accounts of past supervisors and coworkers attesting to the
targets’ outstanding effectiveness, competence, and aggressive achieve-
ment focus. The memo also reported excerpts of the VPs’ personal state-
ments ascribing their success to taking active measures and doing what it
took to succeed in the financial industry.
After reading the memo regarding the first stimulus person, participants
evaluated him or her on a two-page questionnaire. They then read about
and evaluated the second stimulus person and then the third person. The
last of the three managers was always a male target and was included both
to be consistent with the proportions of male and female managers who
would be expected in a male gender-typed position and to make less
apparent our interest in employee sex; data regarding him were not used in
the analyses. The order of presentation of the two managers of interest, one
male and one female, was counterbalanced, and memos about them always
preceded the memo about the third, dummy manager. Also, the information
provided about the first two managers was alternated so that each memo
version was used for the male target half of the time and the female target
the other half. The information regarding the third manager was always the
same, regardless of experimental condition.
Experimental Manipulations
Sex of stimulus person. Each participant was exposed both to a male
and to a female target (as well as to a third target who was male but was
not of interest to us). We manipulated information regarding target sex by
altering the names and gender pronouns in the memos describing each
manager.
Type of information. The type of additional information provided
about the stimulus person was manipulated in the CEO’s introductory
memo. In both the communal and the positive noncommunal information
conditions, two sentences were added at the beginning of the last para-
graph. Because participants reviewed both a male and a female manager,
we had two versions of each information manipulation. In the communal
information condition, the added comments were either
Although Andrea’s [James’s] coworkers agree that she [he] demands
a lot from her [his] employees, they have also described her [him] as
an involved manager who is caring and sensitive to their needs. She
[he] emphasizes the importance of having a supportive work environ-
ment and has been commended for her [his] efforts to promote a
positive community.
or
Subordinates have often described Andrea [James] as someone who is
tough, yet understanding and concerned about others. She [he] is
known to encourage cooperation and helpful behavior, and has
worked hard to increase her [his] employees’ sense of belonging.
In the positive noncommunal information conditions, the added com-
ments were similar in length and design to the comments in the communal
information condition, although the information provided concerned pos-
itive but not communal attributes. They were either
Although Andrea’s [James’s] coworkers agree that she [he] demands
a lot from her [his] employees, they have also described her [him] as
someone who is fair-minded in her [his] treatment of others. She [he]
emphasizes the importance of having an open forum for the exchange
of ideas and has been commended for her [his] efforts to promote
performance excellence within her [his] department.
or
Subordinates have often described Andrea [James] as someone who is
tough, yet outgoing and personable. She [he] is known to reward
individual contributions and has worked hard to maximize her [his]
employees’ creativity.
In the control condition, the CEO’s memo was presented without any
information added to the last paragraph.
Dependent Measures
There were two measures of liking. The first was a composite of
responses to two 9-point scale items (␣⫽.80), a bipolar adjective scale
rating (likable–not likable) and the question “How much do you think you
would like this individual?” (very much–not at all). The second liking
measure, obtained on the second questionnaire after each of the two
stimulus managers had been reviewed separately, was the participants’
choice of which of the two managers they thought they would like better.
We created a perceived interpersonal hostility scale by combining par-
ticipants’ ratings of the target individuals on five 9-point bipolar adjective
scales. The scale items (abrasive–not abrasive, pushy–not pushy,
untrustworthy–trustworthy, manipulative–not manipulative, and selfish–
not selfish; ␣⫽.70) were compiled from those used in previous research
(Heilman et. al., 1989, 1995, 2004).
Boss desirability was assessed by two measures. The first was a single
item 9-point scale asking, “How much would you want this person as your
boss?” (very much–not at all). The second measure, obtained on the second
questionnaire after the two stimulus targets had been reviewed, was the
participant’s choice of which target they would prefer as a boss.
In addition to the dependent variable measures, we created a perceived
communality manipulation check by combining four 9-point bipolar ad-
jective scales: supportive–not supportive, understanding–not understand-
ing, sensitive–insensitive, and caring–not caring (␣⫽.90). We also
checked to make sure that the communality manipulation did not differ-
entially affect the perceived agenticism or competence of the managers.
Our perceived agenticism measure was a composite of six 9-point bipolar
adjective scales: strong–weak, assertive–not assertive, tough–not tough,
bold–timid, active–passive, and dominant–submissive (␣⫽.82). Our per-
ceived competence measure (␣⫽.79) was a composite of three 9-point
bipolar adjective scales (competent–incompetent, effective–ineffective, and
productive– unproductive) and one 9-point question asking, “How success-
ful do you expect this individual to be in the new position?” (very
successful–not successful at all). All the dependent measure scales were
labeled only at the endpoints.
83
PENALTIES FOR WOMEN’S SUCCESS
Results
Manipulation Check
Analysis of variance of the communality ratings indicated a
significant main effect for information, F(2, 72) ⫽ 10.78, p ⬍
.001, indicating that our information manipulation was effective.
Participants perceived managers in the communal condition (M ⫽
5.37) to be more communal than managers for whom no additional
information (M ⫽ 4.23) or positive information that was not
communal (M ⫽ 4.09) was provided. Analysis of variance further
indicated that perceptions of the managers’ agenticism did not vary
as a function of either the sex of the manager or the information
manipulation. Additionally, analysis of variance indicated no sig-
nificant effects in competence perceptions as a function of infor-
mation condition, although women (M ⫽ 7.88) were generally
rated as less competent than men (M ⫽ 8.18), F(1, 72) ⫽ 4.01, p ⬍
.05. Competence in all conditions was thought to be very high,
with the means ranging from 7.78 to 8.19 on a 9-point scale.
Data Analysis
Initial analyses of variance with participant sex included as a
factor indicated no significant main effects or interactions involv-
ing participant sex on any of the three dependent measure scales.
Additionally, the results of the chi-square analyses of the choice
measures did not differ when the data of male and female partic-
ipants were analyzed separately. Consequently, the data of male
and female research participants were combined for all subsequent
analyses.
A multivariate analysis of variance (MANOVA) was conducted
on the two composite scales (likability and perceived interpersonal
hostility) as well as the single-item scale of boss desirability.
Results indicated significant main effects for both target sex, F(3,
70) ⫽ 4.12, p ⬍ .01, and type of information, F(6, 142) ⫽ 2.58,
p ⬍ .05, and a significant Sex of Target ⫻ Information Type
interaction, F(6, 142) ⫽ 3.26, p ⬍ .01. Univariate analyses of
variance were then conducted on each of the scales, and intercell
comparisons were conducted both to clarify interaction effects and
to directly test the hypotheses by examining differences in reac-
tions to male and female managers in each information condition.
Because of the within-subject nature of the sex of manager ma-
nipulation, paired t tests were used to test these differences. The
correlation between the likability and interpersonal hostility rat-
ings was ⫺.55, the correlation between the likability and boss
desirability ratings was .58, and the correlation between the inter-
personal hostility and boss desirability ratings was ⫺.41. Table 1
presents the means and standard deviations for each of the depen-
dent measure scales. Chi-square tests were used to analyze partic-
ipants’ selections of which manager, the man or the woman, they
liked better and which one they preferred as a boss. Table 2
presents the frequencies of the preference item responses in each
experimental condition.
Likability
Scale ratings. Analysis of variance of participants’ ratings on
the likability scale revealed significant main effects for sex of
target, F(1, 72) ⫽ 4.07, p ⬍ .05 (
2
⫽ .05), and for the type of
information provided, F(2, 72) ⫽ 4.80, p ⬍ .05 (
2
⫽ .12). In
addition, a significant Sex of Target ⫻ Type of Information
interaction was obtained, F(2, 72) ⫽ 8.44, p ⬍ .005 (
2
⫽ .19).
Results of the paired comparisons between male and female
targets supported our predictions. In the no added information
condition, participants rated the female manager as significantly
less likable than the male manager, t(24) ⫽ 2.80, p ⬍ .05. The
same pattern was found in the positive noncommunal information
condition, t(24) ⫽ 2.45, p ⬍ .05. However, when communal
information was provided, this pattern was not evident; in fact,
with communal information the pattern was reversed, and partic-
ipants rated the female manager as significantly more likable than
the male manager, t(24) ⫽ 2.80, p ⬍ .05.
Comparative judgments. The results of the chi-square analysis
of participants’ choices of the more likable manager were consis-
tent with the results of the analyses of the liking ratings. The
choice of the female manager was significantly less frequent than
the choice of the male manager in both the no added information
Table 1
Study 1: Means and Standard Deviations for Dependent Measure Scales
Dependent measure
scale
No added
information
Positive
information
Communal
information
MSDMSDMSD
Likability
Male target 6.28
ad
1.24 6.26
ad
1.48 6.30
a
1.28
Female target 5.34
b
1.39 5.38
b
1.78 7.04
d
1.37
Interpersonal hostility
Male target 4.26
ac
1.03 4.70
a
1.10 4.54
a
1.43
Female target 3.42
b
0.95 3.67
bc
1.21 4.90
a
1.19
Boss desirability
Male target 6.20
a
1.35 5.88
ac
2.09 6.04
a
2.17
Female target 5.08
bc
1.58 4.89
b
2.06 6.56
a
2.02
Note. All ratings were done on 9-point scales, and the higher the number, the more favorable the rating (the
more likable, the less interpersonally hostile, the more desirable a boss). n ⫽ 25 in each condition. Means for
each dependent measure grouping that do not share subscripts differ significantly, as indicated by paired t tests
for comparisons between male and female targets and by least significant difference tests for comparisons
between information conditions.
84
HEILMAN AND OKIMOTO
condition,
2
(1, N ⫽ 25) ⫽ 9.68, p ⬍ .01, and the positive
noncommunal information condition,
2
(1, N ⫽ 23) ⫽ 7.04, p ⬍
.01. This was not the case in the communal information condition,
however. In fact, in the communal information condition, female
targets were selected as the more likable manager more often than
the male targets,
2
(1, N ⫽ 25) ⫽ 6.48, p ⬍ .05.
Perceived Interpersonal Hostility
Analysis of variance of the interpersonal hostility adjective
ratings revealed a significant main effect for sex of target, F(1,
72) ⫽ 10.74, p ⬍ .005 (
2
⫽ .13); a significant main effect for
information condition, F(2, 72) ⫽ 5.48, p ⬍ .01 (
2
⫽ .13); and a
significant Sex of Target ⫻ Information Condition interaction,
F(2, 72) ⫽ 8.05, p ⬍ .005 (
2
⫽ .18). Paired comparisons
comparing ratings of the male and female manager in each infor-
mation condition indicated that, as predicted, female managers
were rated as more interpersonally hostile than male managers in
the no added information condition, t(24) ⫽ 3.94, p ⬍ .005, and in
the positive noncommunal information condition, t(24) ⫽ 3.44,
p ⬍ .005, but not in the communal information condition, in which
participants did not rate male and female managers differently in
terms of interpersonal hostility, t(24) ⫽ 1.30, ns.
Boss Desirability
Scale ratings. Analysis of variance of participants’ ratings of
boss desirability revealed a significant main effect for sex of target,
F(1, 72) ⫽ 5.66, p ⬍ .05 (
2
⫽ .07), but not for the type of
information provided, F(2, 72) ⫽ 1.99. A significant effect was
also obtained for the Sex of Target ⫻ Type of Information inter-
action, F(2, 72) ⫽ 5.71, p ⬍ .01 (
2
⫽ .14).
The results of paired comparisons between ratings of male and
female managers also supported our predictions. Participants rated
the female manager as less desirable as a boss than the male
manager in both the no added information condition, t(24) ⫽ 3.02,
p ⬍ .01, and the positive noncommunal information condition,
t(24) ⫽ 2.83, p ⬍ .01. In the communal information condition,
however, ratings of boss desirability did not differ as a function of
the manager’s sex, t(24) ⫽ 1.24, ns.
Comparative judgments. The results of the chi-square analysis
of participants’ boss preferences indicated that, when asked whom
they preferred as a boss, participants chose the male manager
significantly more often than the female manager in the no added
information condition,
2
(1, N ⫽ 25) ⫽ 23.12, p ⬍ .01, and in the
positive noncommunal information condition,
2
(1, N ⫽ 22) ⫽
13.09, p ⬍ .01. However, in the communal information condition
there was no difference in the frequency with which male and
female managers were chosen,
2
(1, N ⫽ 24) ⫽ 0.00.
Discussion
These results both replicate and extend earlier findings. By
demonstrating that women who were depicted as being successful
at male gender-typed jobs were not liked, were viewed as having
very undesirable interpersonal attributes, and were seen as less
desirable as bosses relative to similarly described male managers,
the data replicate the findings of studies that have been reported in
the literature (Heilman et. al., 1989, 2004; Rudman & Glick,
2001). However, the results also go beyond the earlier findings,
demonstrating that the negativity directed against successful
women did not occur when communal information about them was
provided. Thus, the findings lend insight into the process under-
lying these negative reactions.
It is important to note that the suspension of negativity toward
the successful female managers occurred only when communal
information was supplied; other types of commendable informa-
tion did not have the same effect. This suggests that it was indeed
indication of communality, not simply of favorability, that was
responsible for mitigating the negative reactions. These findings
lend support to our contention that the key issue underlying the
negative reactions to women who are successful in male tasks and
roles is the perceived deficiency in their communality. The provi-
sion of communal information softened reactions to women, erad-
icating (and, in the case of liking, even reversing) the negativity
that was so evident when this information was not made available.
However, information of communality is not likely to always
assuage penalties for women’s unconventional success. Only when
the information is taken to be indicative of the woman’s personal
attributes should it have this salutary effect. The proclivity to see
women who are successful in male domains as deficient in com-
munal qualities is likely to be accompanied by resistance to seeing
them otherwise. Thus, if there is ambiguity in the potentially
disconfirming communality information, it is likely to be dis-
counted or dismissed (Fiske & Taylor, 1991). The focus of our
second study is to demonstrate how important it is for the com-
munality information to provide unequivocally disconfirming in-
Table 2
Study 1: Frequencies of Target Choices on Liking and Boss Preferences
Preference
No added
information
Positive
information
Communal
information
Freq. % Freq. % Freq. %
Liking
Male target 18 72 16 70 8 32
Female target 7 28 7 30 17 68
Boss
Male target 21 84 17 77 12 50
Female target 4 16 5 23 12 50
Note. Freq. ⫽ frequency.
85
PENALTIES FOR WOMEN’S SUCCESS
formation about the successful woman’s attributes if it is to coun-
teract the negativity directed at her.
Study 2
The hierarchical structure of organizations often creates ambi-
guity about the actual impetus for behavior. That is, it is sometimes
unclear whether a particular behavior should be attributed to the
actor or to situational constraints, such as company policies or
superiors’ directives. This raises the question of whether the em-
ployee was actually doing what he or she chose to do or was
merely complying with the demands of the work setting. Thus, the
ambiguity about the source of behavior provides an “out” for the
perceiver through attributional rationalization (Heilman & Haynes,
2005) and is likely to limit the degree to which information about
a woman’s communal behavior actually is taken to provide useful
information about her. In situations in which this is the case, the
communal information should not be an effective antidote to the
communality deficits implied by the woman’s success in a male
role.
We therefore expected that when a woman is successful in male
domains, information of her communality will be of little conse-
quence if there is ambiguity about whether she is the impetus for
the communal behavior or whether her communality is merely a
consequence of her efforts to fulfill an organizational responsibil-
ity. To test these ideas, we used the same procedure as that used in
Study 1 but in this case varied the degree of ambiguity about the
source of the manager’s communality.
As in Study 1, we predicted that female managers who had been
successful in male gender-typed jobs would be disliked more, elicit
more negative interpersonal characterizations, and be viewed as
less desirable as a boss than would similarly successful male
managers unless information about their communality was pro-
vided. However, we did not expect information of communality to
have ameliorative effects unless the communal behaviors associ-
ated with the successful woman were clearly and unambiguously a
reflection of who she was and not a possible result of external
demands.
Hypothesis: Communal information will mitigate the dislike,
negative characterizations, and undesirability as a boss of
successful female managers as compared with successful
male managers only when it is clear that the communality
originates from them and is not ambiguous with respect to its
source.
Method
Participants and Design
Participants were 96 male and female undergraduates recruited from an
introductory psychology course who participated in exchange for partial
fulfillment of course experiment participation requirements. The design
wasa2⫻ 3 mixed factorial with sex of the stimulus person as a
within-subject factor and type of additional information provided as a
between-subjects factor. The type of information conditions included a no
information condition and two communal information conditions, one in
which the source of the communal behavior was clearly the stimulus
person (clear communality), and one in which the source of the communal
behavior was ambiguous and potentially externally induced (ambiguous
communality). Participants were randomly assigned to one of the three
information conditions, and all participants reviewed both a male and a
female target.
Procedure
The procedure and stimulus materials were identical to those of Study 1
except for the experimental manipulations described in the following
section.
Type of Additional Information Manipulation
The type of additional information provided about the stimulus person
was again manipulated in the CEO’s memo. In both the clear and the
ambiguous communality conditions, two sentences were added at the
beginning of the last paragraph. In the clear condition, the two sentences
were the same as those presented in the communal information condition
in Study 1. These sentences alluded not only to the manager’s communal
behavior (creating a supportive work environment and building a positive
community or encouraging cooperation and increasing employees’ sense of
belonging) but also to his or her communal attributes (involved, caring, and
sensitive or understanding and concerned about others), thereby implying
that the communal behaviors were a clear result of the manager’s predis-
positions and personal decisions about how to behave as a manager.
However, in the ambiguous communality condition, the manager was
described as performing the same communal behaviors, but either under
the direction of a supervisor or as part of a company-wide initiative,
thereby implying that the communal behavior might have been performed
to fulfill a job responsibility. In particular, in the ambiguous communality
condition one of the two following passages, appearing equally for male
and female managers, was included just preceding the sentence describing
the communal behaviors performed: “In his [her] last year at [unnamed
company] Inc., James [Andrea] worked for a supervisor who was known
for placing high importance on employee relations. Under his [her] boss’s
direction, James [Andrea] . . .” or “In recent years, [unnamed company] has
updated their mission statement, placing greater importance on understand-
ing the concerns of employees. As part of this company-wide initiative,
James [Andrea] ....” As in Study 1, the control condition memo was
presented without any information added to the last paragraph.
Dependent Measures
The dependent measures were the same as those used in the previous
study, including composite measures of likability (␣⫽.83) and interper-
sonal hostility (␣⫽.71) and the single item assessing boss desirability.
Comparative judgments regarding likability and boss preferences were also
collected. Although we could not directly check participants’ perceptions
of whether the source of the target’s communal behavior was clear or
ambiguous without potentially biasing responses, we included a commu-
nality measure to ensure that the participants interpreted our type of
information manipulation as intended. The communality manipulation
check (␣⫽.86) and measures of perceived agenticism (␣⫽.75) and
competence (␣⫽.75) were also the same as those used in Study 1.
Results
Manipulation Check
Although analysis of variance revealed no significant effects for
the communality ratings, planned comparisons tested at the p ⬍
.05 level revealed that participants in the clear communality con-
dition (M ⫽ 5.37) gave significantly higher ratings of target
communality than those in the no information condition (M ⫽
4.23) or the ambiguous communality condition (M ⫽ 4.09). Ex-
amination of the competence scores indicated that the competence
86
HEILMAN AND OKIMOTO
ratings in all conditions were again very high, with the means
ranging from 7.52 to 8.10 on a 9-point scale, and analysis of
variance revealed no significant effects for either perceived agen-
ticism or perceived competence as a function of information con-
dition, gender of target manager, or the interaction between the
two variables.
Data Analysis
Analyses of variance with participant sex included as a factor
revealed only one effect involving participant sex, a main effect for
likability ratings, F(1, 68) ⫽ 5.68, p ⬍ .05, indicating that female
participants generally rated the manager as more likable than did
male participants. There was no evidence of participant sex inter-
acting with our independent variables, and the responses of male
and female participants were combined for subsequent analyses.
Separate chi-square analyses of the male and female participants’
preference data also indicated no differences as a function of
participant sex, so these data also were combined for subsequent
analyses.
A MANOVA conducted on the three dependent measure scales
(likability, interpersonal hostility, and boss desirability) indicated
no significant main effects for target sex, F(6, 67) ⫽ 2.24, or type
of information, F(6, 136) ⫽ 1.04, but a significant Sex of Target ⫻
Information Type interaction, F(6, 136) ⫽ 2.69, p ⬍ .05. Univar-
iate analyses of variance were then conducted on each of the
scales, and paired t tests were conducted to test hypothesized cell
mean differences between ratings of male and female targets in
each information condition. The correlation between the likability
and the interpersonal hostility scales was ⫺.59, the correlation
between the likability and the boss desirability scales was .64, and
the correlation between the interpersonal hostility and the boss
desirability scales was ⫺.35. Table 3 presents the means and
standard deviations for each of the dependent measure scales.
Chi-square tests were used to analyze participants’ responses to the
forced choice questions involving selection of which manager they
liked better and which manager they preferred as a boss. These
frequencies are presented in Table 4.
Likability
Scale ratings. Analysis of variance of participants’ ratings on
the likability scale revealed no significant main effects for the type
of information provided, F(1, 69) ⫽ 2.03. There was, however, a
significant effect for sex of target, F(1, 69) ⫽ 4.37, p ⬍ .05 (
2
⫽
.06), as well as a significant Sex of Target ⫻ Type of Information
interaction, F(2, 69) ⫽ 8.37, p ⬍ .005 (
2
⫽ .20).
Results of the paired comparisons between male and female
targets supported our hypotheses. When no information was pro-
vided, participants rated the female manager as significantly less
likable than the male manager, t(23) ⫽ 2.97, p ⬍ .01, but this was
not the case when the source of communality was clear. In fact, as
in Study 1, the participants in the clear communality condition
rated the female target as significantly more likable than the male
target, t(23) ⫽ 2.17, p ⬍ .05. When the source of communality was
ambiguous, however, it did not have the same ameliorative effect
as when it was clear. When provided with ambiguous communality
information, participants saw the female target manager as less
likable than the male target manager, t(23) ⫽ 2.75, p ⬍ .05.
Comparative judgments. When participants were asked whom
they liked better, the male or the female manager, their responses
supported our predictions. The female manager was selected sig-
nificantly less frequently than the male manager not only in the no
information condition,
2
(1, N ⫽ 24) ⫽ 12.00, p ⬍ .01, but also in
the ambiguous communality condition,
2
(1, N ⫽ 24) ⫽ 7.04, p ⬍
.01. In the clear communality condition, however, there was no
significant difference between the frequency with which the male
and female managers were chosen,
2
(1, N ⫽ 23) ⫽ 0.33.
Perceived Interpersonal Hostility
Analysis of variance of the interpersonal hostility ratings re-
vealed a significant main effect for sex of target, F(1, 69) ⫽ 6.76,
Table 3
Study 2: Means and Standard Deviations for Dependent Measure Scales
Dependent measure
scale
No added
information
Clear communality
information
Ambiguous
communality
information
M SD M SD M SD
Likability
Male target 6.02
acd
1.74 5.92
abc
1.73 6.12
ad
1.28
Female target 5.08
b
1.34 6.60
d
1.27 5.21
bc
1.71
Interpersonal hostility
Male target 4.68
ac
1.58 4.26
ab
1.60 4.85
a
1.14
Female target 3.63
b
0.97 4.69
ac
1.14 4.04
bc
1.03
Boss desirability
Male target 6.04
a
2.12 5.46
abc
2.28 5.71
abc
1.73
Female target 4.87
bc
1.80 5.96
ab
2.05 4.79
c
1.93
Note. All ratings were done on 9-point scales, and the higher the number, the more favorable (the more likable,
the less interpersonally hostile, the more desirable as a boss) the rating. n ⫽ 24 in each condition. Means for each
dependent measure grouping that do not share subscripts differ significantly, as indicated by paired t tests for
comparisons between male and female targets and by least significant difference tests for comparisons between
information conditions.
87
PENALTIES FOR WOMEN’S SUCCESS
p ⬍ .05 (
2
⫽ .09), and a significant Sex of Target ⫻ Information
Condition interaction, F(2, 69) ⫽ 6.35, p ⬍ .005 (
2
⫽ .16). No
significant main effect for information condition was observed,
F(2, 69) ⫽ 0.74.
Paired comparisons comparing ratings of the male and female
manager in each information condition indicated that female man-
agers were rated as more interpersonally hostile than male man-
agers in the no information condition, t(23) ⫽ 3.47, p ⬍ .01, but
no more interpersonally hostile than the male managers in the clear
communality condition, t(23) ⫽ 1.07, ns. Moreover, as we had
predicted, participants in the ambiguous communality condition
responded similarly to participants in the no information condition,
rating female managers as more interpersonally hostile than male
managers, t(23) ⫽ 2.42, p ⬍ .05.
Boss Desirability
Scale ratings. Analysis of variance of participants’ ratings on
the single-item scale assessing boss desirability revealed no sta-
tistically significant main effects for the sex of the target manager,
F(1, 69) ⫽ 3.83, or for the type of information provided, F(2,
69) ⫽ 0.47. However, a significant effect was found for the Sex of
Target ⫻ Type of Information interaction, F(2, 69) ⫽ 3.70, p ⬍
.05 (
2
⫽ .10).
Paired comparisons between male and female targets indicated
that, as predicted, participants rated the female target manager as
significantly less desirable as a boss than the male target manager
when no added information was provided, t(23) ⫽ 2.93, p ⬍ .01,
but not when clear communal information was provided, t(23) ⫽
1.07, ns. Although the difference in ratings of the male and female
managers when the source of the communal information was
ambiguous was not found to be statistically significant at the p ⬍
.05 level, the trend was in the hypothesized direction, t(23) ⫽ 1.74,
p ⫽ .096.
Comparative judgments. The results of the chi-square analysis
of participants’ choices of the more desirable boss were similar to
the comparative ratings regarding likability. As in Study 1, when
asked whom they preferred as a boss, participants preferred the
male manager significantly more often than female manager in the
no information condition,
2
(1, N ⫽ 24) ⫽ 12.00, p ⬍ .01. There
was also a significant difference in selection of the male and
female managers in the ambiguous communality condition,
2
(1,
N ⫽ 24) ⫽ 12.00, p ⬍ .01, in which male managers were again
selected more frequently. In the clear communality condition,
however, there were no differences in the frequencies of male and
female boss choices,
2
(1, N ⫽ 24) ⫽ 3.00.
Discussion
The results of Study 2, like those of Study 1, indicate that the
negativity produced by a woman’s success in a male gender-typed
job can be averted if there is information that the woman has
engaged in communal behavior. However, the results also make
clear that simply providing evidence that a woman has behaved in
a communal manner is insufficient to counteract the detrimental
effects of her success. Her communal behavior must be seen as a
reflection of who she is, not as mere compliance with demands of
the work setting, if it is to dampen the tendency toward negativity
that we have demonstrated in this and the previous study. As the
results indicate, simply verifying that a woman has displayed
communal behavior does not ensure that she will be seen as an
individual who has communal attributes.
However, what types of information lead to inferences of com-
munal attributes? In Studies 1 and 2 we have provided behavioral
information about a woman’s communality and have found that it
can lead to inferences that she is a communal person. However,
information about communality can take forms other than reports
of behavior. To explore the robustness of our findings, it is
important to determine whether nonbehavioral forms of informa-
tion about communality also can attenuate negative reactions to
successful women. This is the focus of Study 3.
Study 3
Lifestyle decisions are an important source of information about
a woman’s communality. In particular, the decision to have chil-
dren is likely to imply communal attributes. Although the infor-
mation it carries about communality is not as explicit as the report
of actual behavior, the role of mother carries with it a set of
inferences about selflessness, caring, and nurturing that are a
central part of the communality construct. In this study, we sought
to determine whether women who are clearly successful in male
domains and also are known to be mothers induce less negative
reactions compared with men than women who also are clearly
successful but do not have children.
Table 4
Study 2: Frequencies of Target Choices on Liking and Boss Preferences
Preference
No added
information
Clear communality
information
Ambiguous
communality
information
Freq. % Freq. % Freq. %
Liking
Male target 18 75 11 46 16 70
Female target 6 25 13 54 7 30
Boss
Male target 18 75 9 38 18 75
Female target 6 25 15 62 6 25
Note. Freq. ⫽ frequency.
88
HEILMAN AND OKIMOTO
Suggesting that motherhood may mitigate negative reactions to
highly successful working women may seem paradoxical because
research evidence has repeatedly shown maternal status to lead to
negative perceptions of a woman’s competence and her commit-
ment to her job (Cuddy, Fiske, & Glick, 2004; Williams & Segal,
2003). However, we are positing that in those instances when a
woman’s success is incontrovertible and her competence and com-
mitment thus undeniable, information of maternal status will be
helpful to her because it verifies her communality.
We therefore expected that when women who had been suc-
cessful in male gender-typed jobs were also mothers, there would
be less of a tendency to react negatively to them as compared with
men than when they were not mothers. We did not expect parental
status to affect reactions to successful men because the boost in
communality perceptions that we expected to accompany parental
status should not, any more than our communal information ma-
nipulations in the first two studies, be of consequence in their
evaluations. Study 3 used the same research procedure as the first
two studies, except that parental status was systematically varied.
The dependent measures again included composite scales of lik-
ability, perceptions of interpersonal hostility, and the single-item
assessment of desirability as a boss. We formulated the following
hypothesis:
Hypothesis: Female managers who are successful in male
gender-typed jobs will be disliked more, will elicit more
negative interpersonal characterizations, and will be found
less desirable as bosses than similarly successful male man-
agers when information about parental status is not provided
rather than when information about parental status is
provided.
Method
Participants and Design
Participants included 46 male and female undergraduates recruited from
an introductory psychology course who participated in exchange for partial
fulfillment of course experimental participation requirements. The design
wasa2⫻ 2 mixed factorial with sex of the stimulus person (male or
female) a within-subject factor and parental status information (no parental
status information or information that the target has children) a between-
subjects factor. Participants were randomly assigned to one of the parental
status information conditions, and participants reviewed both a male and a
female target.
Procedure
The procedure and stimulus materials were the same as those used in
Studies 1 and 2 except for the parental status manipulation, which is
described in the following section.
Parental Status Manipulation
The CEO’s letter from the control conditions of Studies 1 and 2 was used
in all conditions of Study 3. The parental status information was manipu-
lated by the addition of a single sentence at the beginning of the last
paragraph. In the condition that provided information that the manager had
children, the additional sentence indicated that he or she was moving to the
city with his or her children. Because respondents evaluated both a male
and a female target manager, there were two versions of this statement,
each assigned with equal frequency to the male and female manager. They
read as follows: “James [Andrea] plans to move to Chicago with his [her]
two children in the next few weeks, and looks forward to meeting his [her]
new staff” or “James [Andrea], his [her] son, and his [her] daughter, will
be moving to the area in the near future.” The no information about
parental status condition memos simply left out mention of children. In all
cases, there was indication that the manager was married.
Dependent Measures
The dependent measures were identical to those used in the previous two
studies, including composite measures of likability (␣⫽.81) and inter-
personal hostility (␣⫽.71) and the single-item desirability as boss mea-
sure. Comparative judgments regarding likability and boss preferences
were also collected. Although we could not directly ask participants to
report whether the target managers had children without signaling our
interest in the parenthood variable, we again included a communality
measure to ensure that the participants interpreted our type of information
manipulation as intended. Using the same measures as in Studies 1 and 2,
we collected ratings of communality (␣⫽.90) as well as agenticism (␣⫽
.79) and competence (␣⫽.85).
Results
Manipulation Checks
Analysis of variance of ratings on the communal scale indicated
that our parental status manipulation had its intended effect. There
was a significant main effect for parental status information, F(2,
66) ⫽ 3.45, p ⬍ .05, with managers who were described as having
children (M ⫽ 5.71) viewed as more communal than managers for
whom no information regarding parental status was provided (M ⫽
4.83). Analysis of variance also indicated that there were no
significant effects of our independent variables on agenticism
ratings or competence ratings and that, as in the earlier studies,
managers in all conditions were viewed as highly competent, with
cell means ranging from 7.68 to 8.27 on a 9-point scale.
Data Analysis
Initial analyses revealed a significant main effect for participant
sex on the boss desirability ratings, F(1, 43) ⫽ 6.67, p ⬍ .05,
indicating that women were generally more favorable in their
ratings than were men. However, there were no interaction effects
involving participant sex, which indicated that the pattern of re-
sponses of men and women as a function of our independent
variables did not differ, and therefore the responses of male and
female participants were combined for subsequent analyses. Sep-
arate analyses of the choice data provided by male and female
participants also showed no differences in pattern as a function of
sex of the respondent, so the data of male and female participants
were combined for subsequent chi-square analyses.
A MANOVA conducted on the three dependent measure scales
(likability, interpersonal hostility, and boss desirability) indicated
no significant main effects for parental status information, F(3,
42) ⫽ 1.96, or sex of target, F(3, 42) ⫽ 1.28, but a significant
effect was found for the Sex of Target ⫻ Parental Status interac-
tion, F(3, 42) ⫽ 4.62, p ⬍ .01. Univariate analyses of variance
were then conducted on each of the scales, and paired t tests were
conducted to test the predicted differences between ratings of male
and female targets in each information condition. The correlation
between the ratings on the likability and interpersonal hostility
89
PENALTIES FOR WOMEN’S SUCCESS
scales was ⫺.54, the correlation between the likability and the boss
desirability scales was .81, and the correlation between the inter-
personal hostility and the boss desirability scales was ⫺.53. Table
5 presents the means and standard deviations for each of the
dependent measure scales. Chi-square tests were used to analyze
participants’ responses to the forced choice questions involving
selection of which manager they liked better and which they
preferred as a boss. These frequencies can be found in Table 6.
Likability
Scale ratings. Analysis of variance of participants’ ratings on
the likability scale revealed no significant main effects for sex of
target, F(1, 44) ⫽ 2.76. There was, however, a significant effect
for parental status, F(1, 44) ⫽ 4.49, p ⬍ .05 (
2
⫽ .09), as well as
a significant Sex of Target ⫻ Parental Status interaction, F(1,
44) ⫽ 8.44, p ⬍ .01 (
2
⫽ .16).
Paired comparisons between ratings of male and female man-
agers supported our hypotheses. They indicated that when no
parental information was provided, participants rated the female
target manager as significantly less likable than the male target
manager, t(22) ⫽ 3.00, p ⬍ .01, but when information was
provided that the manager had children, there was no difference in
the likability ratings of the male and female manager, t(22) ⫽ 0.96.
Comparative judgments. The results of the chi-square analysis
of participants’ choice of the more likable manager were consistent
with the results of the analyses of the liking ratings. In the no
information condition, the choice of the female manager was
significantly less frequent than the choice of the male manager,
2
(1, N ⫽ 22) ⫽ 5.82, p ⬍ .05. However, when the manager was
said to have children, there was no significant difference in the
frequency of the choice of the male and female managers,
2
(1,
N ⫽ 23) ⫽ 0.78.
Perceived Interpersonal Hostility
Analysis of variance of the interpersonal hostility adjective
ratings revealed no significant main effects for sex of target, F(1,
44) ⫽ 0.40. Again, a significant main effect was found for parental
status information, F(1, 44) ⫽ 4.37, p ⬍ .05 (
2
⫽ .09), as well as
a significant Sex of Target ⫻ Parental Status Information interac-
tion, F(1, 44) ⫽ 7.22, p ⬍ .01 (
2
⫽ .14).
Paired comparisons of the ratings of the male and female man-
ager in each parental status condition indicated that, as anticipated,
when there was no indication of parental status, female managers
were rated as more interpersonally hostile than male managers,
t(22) ⫽ 2.47, p ⬍ .05, but when the managers were said to have
children, the female managers were rated no differently in terms of
interpersonal hostility than the male managers, t(22) ⫽ 1.39, ns.
Boss Desirability
Scale ratings. Although the means fell in the predicted pattern
(see Table 5), analysis of variance of participants’ ratings on the
boss desirability measure revealed no significant main effects for
sex of manager, F(1, 44) ⫽ 0.382, and no significant interaction
effect, F(1, 44) ⫽ 1.362. There was, however, a significant main
effect for parental status information, F(1, 44) ⫽ 4.320, p ⬍ .05
(
2
⫽ .09), with parents rated as generally less desirable as bosses.
Comparative judgments. Despite the lack of support for our
hypothesis about boss desirability in our rating measure, the forced
choice boss preferences fell into the predicted pattern. When
participants were asked whom they preferred as a boss, the choice
of the female manager was significantly less frequent than the
choice of the male manager in the no information condition,
2
(1,
N ⫽ 22) ⫽ 4.26, p ⬍ .01. However, when managers were thought
to have children, frequencies of choices between male and female
managers did not differ from one another,
2
(1, N ⫽ 23) ⫽ 0.78, ns.
Discussion
The results were consistent with those of the first two studies. In
the absence of information about parental status, successful female
managers were viewed far more negatively than identically de-
scribed male managers. However, as hypothesized, information
that the successful female manager was a mother, which created
perceptions of her as a communal person, eliminated the negativity
Table 5
Study 3: Means and Standard Deviations for Dependent
Measure Scales
Dependent measure
scale
No parental
status information
Target has
children
MSDMSD
Likability
Male target 6.26
a
1.34 6.41
a
1.22
Female target 5.30
b
1.80 6.67
a
1.19
Interpersonal hostility
Male target 4.42
a
1.30 4.38
a
1.38
Female target 3.64
b
0.85 4.86
a
1.37
Boss desirability
Male target 5.52
ab
1.93 6.04
a
1.52
Female target 4.96
b
2.18 6.22
a
1.51
Note. All ratings were done on 9-point scales, and the higher the number,
the more favorable (the more likable, the less interpersonally hostile, the
more desirable as a boss) the rating. n ⫽ 24 in each condition. Means for
each dependent measure grouping that do not share subscripts differ
significantly, as indicated by paired t tests for comparisons between male
and female targets, and by least significant difference tests for comparisons
between parental status conditions.
Table 6
Study 3: Frequencies of Target Choices on Liking and Boss
Preferences
Preference
No parental
status information
Target has
children
Freq. % Freq. %
Liking
Male target 15 68 10 43
Female target 7 32 13 57
Boss
Male target 14 64 10 43
Female target 8 36 13 57
Note. Freq. ⫽ frequency.
90
HEILMAN AND OKIMOTO
directed at her. These data therefore lend additional support to the
idea that the perceived deficit in communality drives negative
reactions to women who are successful at traditionally male jobs.
They also suggest that although motherhood has been shown to be
so problematic for women seeking to advance their careers, when
there is a perceived communality deficit, as appears to be the case
with successful female managers, being a mother can be advanta-
geous rather than disadvantageous for career-oriented women.
General Discussion
The studies presented in this article replicate the findings of
others, showing that women who are successful in male gender-
typed domains are penalized for their success. They indicate that
these women are decidedly more disliked and interpersonally
derogated than identically described men and, in addition, that they
are found to be less desirable as bosses. However, the studies also
indicate that information that causes such women to be seen as
communal appears to counteract these negative responses. The
information provided must be of the sort that not just is favorable
but also affects the perceptions of the successful woman’s com-
munal attributes (Study 1) and does not merely report the perform-
ing of communal behavior without verifying that the behavior’s
source is the woman herself (Study 2). Moreover, the information
of communality need not be explicitly behavioral; it can be con-
veyed by a role inferred to require communal traits, such as
motherhood (Study 3).
These results lend support to the idea that when women are
successful in male domains, it is their violation of gender-
stereotypic prescriptions that produces negative reactions to them.
Their success and the agenticism that it entails apparently cause
them to be seen as deficient in the feminine attributes mandated by
gender stereotypes, and they are penalized as a result. Because the
negative reactions to successful women are alleviated with evi-
dence that they are, in fact, communal, our data suggest that these
women’s perceived violation of feminine “shoulds,” not their
taking on of masculine “should nots,” underlies and fuels the
penalties these women incur for their success.
Furthermore, although our results suggest that women who are
believed to have behaved agentically are presumed to have vio-
lated prescriptive injunctions to be communal, they do not imply
that this inference is inevitable or immutable. On the contrary, our
results indicate that it is possible for women who are successful to
be seen as both agentic and communal simultaneously. This find-
ing is important and challenges the idea that being respected for
task competence and being liked for interpersonal qualities are
mutually exclusive reactions to individuals from stereotyped
groups (Glick & Fiske, 1997).
It is important to point out once again that information of a
woman’s communality can be a double-edged sword. There is a
great deal of literature attesting to the deleterious consequences of
women being seen as communal rather than agentic because of the
presumption of incompetence at supposedly male tasks that this
stereotyped perception evokes (Heilman, 1995, 2001). Yet the
research reported in this study suggests that when women’s com-
petence is beyond question, perceptions of communality may
sometimes work in their favor, counteracting negative interper-
sonal responses and perhaps curtailing career-hindering reactions
that have been found to ensue (Heilman et al., 2004). However,
because our dependent variables did not include actual personnel
decisions, it is premature to make claims about possible effects on
career advancement; it is conceivable that the favorable affective
reactions are not accompanied by similarly favorable employment
decisions. Needless to say, this potentially paradoxical effect of
communality on evaluations of women is worthy of more study, as
is the question of what types of information effectively commu-
nicate that a woman is communal.
It is interesting to note that although female participants occa-
sionally demonstrated a tendency to be more lenient in their ratings
than male participants, in none of our studies did male and female
participants respond differently to the male or to the female man-
ager as a consequence of the experimental manipulations. This
finding, which is consistent with that of earlier studies (Heilman et
al., 2004), suggests that women and men subscribe to the same
normative gender prescriptions and enforce them in similar ways.
It also attests to the power and universality of the prescriptive
aspects of gender stereotypes in our culture.
In both Study 1 and Study 2, there was an apparent augmenta-
tion in the likability ratings of the successful women about whom
unambiguous communal information was provided. In contrast to
the negative reactions to the women about whom no communal
information was supplied, participants rated the communal female
manager as even more likable than her male counterpart. This bias
toward favorability might have been brought about by the sharp
contrast between observation and stereotypical expectations (Feld-
man, 1981; Weber & Crocker, 1983), by the perception that our
successful women who also were communal were truly extraordi-
nary, or, more simply, by the belief that women who are norma-
tively appropriate and have feminine attributes are inherently more
likable than men—the “women are wonderful” effect reported by
Eagly and Mladinic (1989; Eagly, Mladinic, & Otto, 1991). These
possible explanations can only be disentangled by future research.
Further research should also address the question of whether
men, too, are penalized for success that violates gender norms. If
they are, which we suspect to be the case, it would be because they
violate gender prescriptions to be agentic. Therefore, the negative
consequences of disapproval are likely to take a different form
than those for women who violate gender norms. Moreover, the
type of information needed to combat the negativity also might
differ. Research addressing these questions is currently underway.
There are limitations of the present research that should be
noted. First, the use of paper-and-pencil stimulus materials in these
three studies, although it allowed us to test the role of prescriptive
stereotypes in a controlled manner, limits the degree to which
conclusions can be drawn about how people react to successful
women in actual work situations. There is no question that partic-
ipation in the research session lacked the richness and intricacies
of involvement in an actual organizational setting and in true-life
relationships and that the questions we address in this study need
to be explored further in a field setting where work relationships
are of consequence and more textured information about cowork-
ers is readily available. Moreover, because our participants were
put into the role of potential subordinates, it is important to
determine whether there are similar reactions on the part of those
who are organizationally senior to the successful female manager.
In addition, although many of the undergraduate participants were
soon to be entry-level employees themselves and although assess-
ment of the participant pool indicated that 91% of them had
91
PENALTIES FOR WOMEN’S SUCCESS
worked for more than a 1-year period (with an average of 3.4
years), they still might not have had the organizational experience
that would ensure that their responses were representative of
people in work settings. Therefore, it is also important to test out
these ideas using a sample of working people. Thus, although we
have demonstrated that the penalties women incur for being suc-
cessful in male-dominated areas can, under some circumstances,
be mitigated by communality information, specification of when,
for whom, and under what conditions this occurs remains to be
determined.
Nevertheless, the potential implications of our results for
women in the workplace are highly provocative. They suggest that
if the assumptions arising from women’s success in male-
dominated fields are counteracted with evidence of communality,
negativity directed at these women can be averted. Thus, the
results may help explain why some women are not burdened with
the pernicious and potentially career-limiting consequences of
treading where they are believed not to belong. However, the
results also starkly highlight that the path up the organizational
ladder can be replete with potential minefields for women who are
striving to reach the upper echelons of power and authority. Doing
what men do, as well as they do it, does not seem to be enough;
women must additionally be able to manage the delicate balance of
being both competent and communal.
References
Bartol, K. M., & Butterfield, D. A. (1976). Sex effects in evaluating
leaders. Journal of Applied Psychology, 61, 446 – 454.
Burgess, D., & Borgida, E. (1999). Who women are, who women should
be: Descriptive and prescriptive gender stereotyping in sex discrimina-
tion. Psychology, Public Policy, and Law, 5, 665– 692.
Carli, L. L. (1990). Gender, language, and influence. Journal of Person-
ality and Social Psychology, 59, 941–951.
Carli, L. L., LaFleur, S. J., & Loeber, C. C. (1995). Nonverbal behavior,
gender, and influence. Journal of Personality and Social Psychology, 68,
1030 –1041.
Cejka, M. A., & Eagly, A. H. (1999). Gender-stereotypic images of
occupations correspond to the sex segregation of employment. Person-
ality and Social Psychology Bulletin, 25, 413– 423.
Cialdini, R. B., & Trost, M. R. (1998). Social influence: Social norms,
conformity, and compliance. In D. T. Gilbert, S. T. Fiske, & G. Lindzey
(Eds.), The handbook of social psychology (4th ed., Vol. 2, pp. 151–
192). Boston: McGraw-Hill.
Cuddy, A. J. C., Fiske, S. T., & Glick, P. (2004). Working moms can’t win:
Mothers are incompetent and female professionals are cold. Journal of
Social Issues, 60, 701–718.
Dipboye, R. L. (1985). Some neglected variables in research on discrim-
ination in appraisals. Academy of Management Review, 10, 116–127.
Eagly, A. H., & Karau, S. J. (2002). Role congruity theory of prejudice
toward female leaders. Psychological Review, 109, 573–598.
Eagly, A. H., & Mladinic, A. (1989). Gender stereotypes and attitudes
toward women and men. Personality and Social Psychology Bulletin, 15,
543–558.
Eagly, A. H., Mladinic, A., & Otto, S. (1991). Are women evaluated more
favorably than men? An analysis of attitudes, beliefs, and emotions.
Psychology of Women Quarterly, 15, 203–216.
Feldman, J. M. (1981). Beyond attribution theory: Cognitive processes in
performance appraisal. Journal of Applied Psychology, 66, 127–148.
Fiske, S. T., & Taylor, S. E. (1991). Social cognition (2nd ed.). New York:
McGraw-Hill.
Glick, P., & Fiske, S. T. (1997). Hostile and benevolent sexism: Measuring
ambivalent sexist attitudes toward women. Psychology of Women Quar-
terly, 21, 119–135.
Heilman, M. E. (1983). Sex bias in work settings: The lack of fit model. In
B. Staw & L. Cummings (Eds.), Research in organizational behavior
(Vol. 5, pp. 269 –298). Greenwich, CT: JAI Press.
Heilman, M. E. (1995). Sex stereotypes and their effects in the workplace:
What we know and what we don’t know. Journal of Social Behavior and
Personality, 10, 3–26.
Heilman, M. E. (2001). Description and prescription: How gender stereo-
types prevent women’s ascent up the organizational ladder. Journal of
Social Issues, 57, 657– 674.
Heilman, M. E., Block, C. J., & Martell, R. (1995). Sex stereotypes: Do
they influence perceptions of managers? Journal of Social Behavior and
Personality, 10, 237–252.
Heilman, M. E., Block, C., Martell, R., & Simon, M. (1989). Has anything
changed? Current characterizations of men, women, and managers.
Journal of Applied Psychology, 74, 935–942.
Heilman, M. E., & Haynes, M. C. (2005). No credit where credit is due:
Attributional rationalization of women’s success in male–female teams:.
Journal of Applied Psychology, 90, 905–916.
Heilman, M. E., Martell, R., & Simon, M. (1988). The vagaries of sex bias:
Conditions regulating the undervaluation, equivaluation, and overvalu-
ation of female job applicants. Organizational Behavior and Human
Decision Processes, 41, 98 –110.
Heilman, M. E., Wallen, A. S., Fuchs, D., & Tamkins, M. M. (2004).
Penalties for success: Reactions to women who succeed at male tasks.
Journal of Applied Psychology, 89, 416–427.
Jago, A. G., & Vroom, V. H. (1982). Sex differences in the incidence and
evaluation of participative leader behavior. Journal of Applied Psychol-
ogy, 67, 776–783.
Nieva, V. F., & Gutek, B. A. (1980). Sex effects on evaluation. Academy
of Management Review, 5, 267–276.
Powell, G. N., Butterfield, D. A., & Parent, J. D. (2002). Gender and
managerial stereotypes: Have the times changed? Journal of Manage-
ment, 28, 177–193.
Rudman, L. A. (1998). Self-promotion as a risk factor for women: The
costs and benefits of counterstereotypical impression management. Jour-
nal of Personality and Social Psychology, 74, 629 – 645.
Rudman, L. A., & Glick, P. (2001). Prescriptive gender stereotypes and
backlash toward agentic women. Journal of Social Issues, 57, 743–762.
Schein, V. E. (2001). A global look at psychological barriers to women’s
progress in management. Journal of Social Issues, 57, 675–688.
Swim, J. K., Borgida, E., Maruyama, G., & Myers, D. G. (1989). Joan
McKay versus John McKay: Do gender stereotypes bias evaluations?
Psychological Bulletin, 105, 409 – 429.
Tosi, H. L., & Einbinder, S. W. (1985). The effects of the type and amount
of information in sex discrimination research: A meta-analysis. Academy
of Management Journal, 28, 712–723.
Weber, R., & Crocker, J. (1983). Cognitive processes in the revision of
stereotypical beliefs. Journal of Personality and Social Psychology, 45,
961–977.
Williams, J. C., & Segal, N. (2003). Beyond the maternal wall: Relief for
family caregivers who are discriminated against on the job. Harvard
Women’s Law Journal, 26, 77–162.
Received April 25, 2005
Revision received January 20, 2006
Accepted January 24, 2006 䡲
92
HEILMAN AND OKIMOTO
