Applying the Lack of Fit Model to Reduce Gender Discrimination

This report examines gender discrimination in the workplace, highlighting the persistence of gender inequalities and the significant role of gender bias. It introduces the lack of fit model as a framework for understanding how gender stereotypes and perceived job requirements create incongruity, leading to negative expectations about women's performance and biased employment decisions. The report explores intervention strategies to reduce these perceptions, including eliminating stereotype-based characterizations and breaking the link between expectations and biased evaluations. It discusses the importance of addressing the factors that drive gender discrimination, with a focus on organizational interventions to promote fair and equitable practices, providing insights into their potential impact and effectiveness. The report emphasizes the need for vigilance in monitoring the effects of these strategies and provides a theoretical context for understanding gender bias in evaluative decision-making.

https://doi.org/10.1177/1368430218761587
Group Processes & Intergroup Relation
2018, Vol. 21(5) 725–744
© The Author(s) 2018
Reprints and permissions:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1368430218761587
journals.sagepub.com/home/gpi
G
P
I
R
Group Processes &
Intergroup Relations
The past 50 years have been marked by consider-
able progressfor womenin the workplace,
reflected in increased labor force participation
rates and a shrinking wage gap (U.S. Bureau of
Labor Statistics, 2017). Yet, these advances can
sometimes be deceptive in their implications.
Although a holistic view of organizations might
suggest that gender parity has been achieved, a
segmented breakdown of professions, positions,
and industries offers a far less favorable picture.
Today, women remain underrepresented in many
prestigious and high-status jobs. They comprise
only 5.8% of S&P 500 CEOs (Catalyst, 2017)
and, in the United States, represent less than 20%
of technicalrolesin majortech companies
(Mundy, 2017) and 11% of tenured professors in
engineering (National Science Foundation, 2015).
Changes in the wage gap, too, have stalled (U.
Bureau of Labor Statistics, 2017).
Although there are many possible contrib
tors to the perpetuation of gender inequality i
the workplace, research suggests that gender
criminationplaysa significantrole (Heilman,
2012). Organizations appear motivated to com
bat gender discrimination, with many publi
carrying out efforts to increase gender equ
Open about their lagging numbers, for exampl
Combatting gender discrimination:
A lack of fit framework
Madeline E. Heilman1 and Suzette Caleo2
Abstract
Gender inequalities in the workplace persist, and scholars point to gender discriminat
contributor. As organizations attempt to address this problem, we argue that theory c
on potential solutions. This paper discusses how the lack of fit model can be used by o
a framework to understand the process that facilitates gender discrimination in emplo
and to identify intervention strategies to combat it. We describe two sets of strategies
at reducing the perception that women are not suited for male-typed positions. The se
preventing the negative performance expectations that derive from this perception of
influencing evaluative judgments. Also included is a discussion of several unintentiona
that may follow from enacting these strategies. We conclude by arguing for the impor
interplay between theory and practice in targeting gender discrimination in the workp
Keywords
gender bias, gender discrimination, gender stereotypes, lack of fit
Paper received 11 May 2017; revised version accepted 2 February 2018
1New York University
2Louisiana State University
Corresponding author:
Madeline Heilman, Department of Psychology, New York
University, 6 Washington Place, New York, NY 10003, US
Email: [e-mail]
761587GPI0010.1177/1368430218761587Group Processe s & Intergroup Relations Heilman and Caleo
research-article2018
Article

726 Group Processes & Intergroup Relations
male-dominated companies in Silicon Valley have
recently pledged to dedicate millions of dollars to
efforts intended to create more inclusive climates
and improve the representation of women in
technology (Mundy, 2017). Yet, these attempts
are proving to be slow going (Isaac, 2015; Wells,
2016). Although more time is needed to deter-
mine if they have had their intended effect, it is
clear that overcoming discrimination is not a sim-
ple task.
Taking a proactive approach to combatting
genderdiscriminationis laudable.However,
intention does not always translate into effective-
ness. If organizations are to effectively address
the shortage of women in male-dominated posi-
tions, the interventions they design have the best
likelihood of succeeding if they take into account
the factors that drive gender discrimination and
target them directly. To this end, this paper draws
from the tenets of the lack of fit model (Heilman,
1983, 2001, 2012) to provide a framework for
identifying strategies for curtailing the effects of
gender stereotyping and gender bias on employ-
ment decisions.
We draw on the lack of fit model for several
reasons. First, it has been specifically tailored to
understand gender discrimination, which is argu-
ably distinct from other forms of discrimination
(Paluck & Green, 2009). Second, the ideas posed
by the model are widely studied and supported by
decades of research (Heilman, 2012), offering
assurance that the strategies we propose are both
theoretically and empirically grounded. Finally,
working from a single theory allows for a system-
atic approach to the design of organizational
intervention strategies. As we will discuss, the
lack of fit model specifies a process underlying
gender bias in evaluative decision making and is
suggestive of various conditions that determine
whether or not this process takes hold. It there-
fore enables a consideration of why and when
discriminatoryemploymentdecisionsoccur,
pointing to contributory factors that can become
the focal point for intervention.
We begin with an overview of the lack of fit
model and its key propositions, later describing
these propositions in more detail as we present
intervention strategies that derive from them. W
will review interventions that already are part o
the organizational arsenal and also propose new
interventions that emerge from a systematic co
sideration of the theory. It is our hope tha
putting these interventions into a theoretical co
text we will provide insight into what they migh
accomplish and when they will be most effectiv
expanding the choices open to organizations
Lastly, we discuss unintended consequences th
could arise from these interventions, emphasizi
the need for vigilance in monitoring their effect
It is of note that our focus is on organization
as the instruments of change. We identify inter
ventions in which organizations can engage tha
impede the processes that perpetuate gender b
and give rise to gender discrimination, not thos
aimed at encouraging women themselves to con
front and overcome the biases they encount
Furthermore, the strategies we propose seek t
address gender discrimination in the formal eva
uative decisions made within organizations,
in the impromptu daily interactions and ass
ments that no doubt also affect women’s career
prospects.
The Lack of Fit Model
Almost 35 years have elapsed since the introdu
tion of the lack of fit model (Heilman, 1983), b
its propositionscontinueto ring true today.
According to the model, outcomes that are dis-
criminatory against women stem from a mis
match between the attributes that women a
thought to possess and the attributes seen as n
essary for success in male-typed positions a
fields. This resulting incongruity forms the basi
of negative expectations about women’s perfor
mance, which thereby bias the processing o
information and, consequently, facilitate discrim
natory behavior.
Central to the model is a consideration of gen
der stereotypes—preconceptions regarding w
men and women are like. Gender stereotypes p
tray men as agentic and women as commun
(Haines, Deaux, & Lofaro, 2016). Whereas m
are thought to be assertive, bold, and aggressiv

Heilman and Caleo 727
women are thought to be relationship-oriented,
nurturing, and kind. These stereotypical depictions
also tend to be oppositional, with both women and
men viewed as lacking what is thought to be most
prevalent in the other sex. Thus, women are seen
not only as communal but also as not agentic.
Stereotypic beliefs about men and women are both
widely and consistently held, and have proven to
be resistant to change despite decades of social
progress (Haines et al., 2016).
Genderstereotypescreatedifficultiesfor
womenpursuingtraditionallymale work,as
femalestereotypesare incompatiblewith the
attributes and behaviors thought necessary for
successin thosefields.In additionto being
numerically dominated by men, these male-typed
fields and occupations are conceived of as neces-
sitating stereotypically masculine attributes (Cejka
& Eagly, 1999; Johnson, Murphy, Zewdie, &
Reichard, 2008). Thus, in contrast to men, who
likely are advantaged by the overlap between per-
ceivedattributesand job requirements (Kray,
Galinsky, & Thompson, 2002; Ryan et al., 2016),
women are disadvantaged. The inconsistency, or
lack of fit, between stereotypes about women and
the perceived requirements for success in male-
typedpositionsleadsto the perceptionthat
women are not well suited for them, producing
negative expectations about their likely perfor-
mance. These expectations in turn lead to the
presumption that women lack the competence
necessary to do well in these positions and that
they are unlikely to succeed (Heilman, 2012).
Stereotype-based expectations about women’s
competence can bias the ways in which individu-
als make decisions about applicants and employ-
ees. They are self-perpetuating,skewing
information processing in a direction that vali-
dates decision makers’ initial stereotypic beliefs.
Thus, negativeperformanceexpectationscan
impact the kinds of information that evaluators
recall (Hollingshead & Fraidin, 2003; Martell,
1996) and attendto (Biernat,Fuegen,&
Kobrynowicz,2010),and can influencehow
informationis interpreted(Madera,Hebl, &
Martin, 2009). They also can affect how individu-
als weight and reconcile different sources and
typesof information(Norton,Vandello,&
Darley, 2004). These effects on information pr
cessing are fueled by ambiguity in the evaluat
process, which increases reliance on expectati
and thereforefacilitatescognitivedistortion
(Heilman & Haynes, 2008).
Because information processing is funda-
mental to a range of human resource manage
ment decisions (Murphy & Cleveland, 1995
the negativeperformanceexpectationsthat
arise from the perception of lack of fit betwee
what women are like and what is required
perform in a male-typed position are likely
promotegenderbias in evaluativedecision
making and prompt discrimination in recru
ment (Gaucher, Friesen, & Kay, 2011), selectio
(Biernat & Fuegen, 2001), performance
appraisal (Bauer & Baltes, 2002), promotio
(Lyness & Heilman, 2006), and compensati
(Castilla, 2008). Figure 1 provides a depict
of the lack of fit model and its effects.
Given the persistence of gender stereotype
and the tenacity of negative performance exp
tations about women in male-typed positions,
is unsurprising to see that gender equality in t
workplace has not yet been achieved. Yet, ove
coming gender discrimination is not impossibl
Organizations hoping to address gender dis
crimination must directly target the factors th
give rise to it and the conditions that promote
The remaining portion of the paper is dedicat
to reviewingand proposingstrategies,built
upon the tenets of the lack of fit model, that d
just that.
Strategies
We broadly identify two sets of intervention st
gies. The first involves decreasing lack of fit pe
tions. The second involves breaking the link be
the expectations of incompetence that arise fr
lack of fit perceptions and the gender bias in e
ative decision making it fosters. Although the
vention strategies are presented individually,
not necessarily independent; under some cond
they may stand on their own, but under others
may be more effective if used in combinat

728 Group Processes & Intergroup Relations
Figure 2 depicts the gender bias process and these
leverage points for intervention.
Decreasing Lack of Fit Perceptions
Research indicates that the greater the perceived
lack of fit between the two components of the
lack of fit model—gender stereotype-based cha
acterizations and gendered job requirements
the more negative the resulting bias-produci
expectations. The lack of fit model suggests tha
minimizing either component of the model will
abate perceptions of incongruity and thereb
prevent negative performance expectations a
Figure 1. The lack of fit model.
Figure 2. The gender bias process and leverage points for intervention.

Heilman and Caleo 729
biased decision-making processes from taking
hold. We discuss each of these below.
Eliminating Stereotype-Based Characterizations. Detri-
mental stereotype-based characterizations can be
eliminated from decision making by changing ste-
reotypical beliefs about women or by inhibiting
their use by reducing the salience of gender.
Changinggenderstereotypes. Thelack of fit
model is grounded on the premise that gender
stereotypes dominate in the workplace, shaping
the ways applicants and employees are perceived.
This suggests that in efforts to curb gender dis-
crimination, organizations should focus on elimi-
nating gender stereotypes. This seems obvious.
In fact, research has shown that people who
do not hold stereotypic beliefs are less likely to
engage in gender discriminatory behavior (Bauer
& Baltes, 2002). However, accomplishing this is
neither easy nor straightforward. Despite dec-
ades of concerted efforts by organizations to
eliminate stereotypes—most evident in training
and other types of educational programs (Kulik
& Roberson, 2008)—the results are not encour-
aging. Studies suggest that stereotypes have not
changed appreciably (Haines et al., 2016), and
the incidence of gender discrimination has not
decidedly declined (Jones, Peddie, Gilrane, King,
& Gray, 2016; Koch, D’Mello, & Sackett, 2015).
Such findings suggest that it is more difficult to
change minds than organizations might believe.
Educationand training. Theremedymost
frequently employed by organizations to address
discrimination, diversity training, is designed to
confront decision makers’ biases and to lessen the
propensity to stereotype (Dobbin, Kalev, & Kelly,
2007). Yet, the evidence surrounding its effective-
ness is at best mixed, appearing to depend on the
outcome in question. According to a recent meta-
analysis (Bezrukova, Spell, Perry, & Jehn, 2016),
the cognitive knowledge obtained during diver-
sity training tends to endure; however, changes
in attitudinal outcomes tend to fade over time.
Evidence relating to behavioral outcomes also is
not encouraging (Bezrukova et al., 2016). Kalev,
Dobbin, and Kelly (2006), for example, found
diversity training to be among the least effecti
diversity management techniques, noting th
does little to increase the proportion of wome
and members of minority groups in manageri
positions.
Despite this unimpressive record of success
perhaps because of it, alternative forms of trai
continue to emerge, the most recent being unc
scious bias training (Lublin, 2014). This ty
training involves making people aware of impl
stereotypes and biases, with an eye on reducin
automaticassociationsthey may hold. Some
research suggests positive outcomes (Carnes e
2015; Jackson, Hillard, & Schneider, 2014), bu
is little evidence of its actual effectiveness in q
the use of stereotypes in workplace setting
short, despite the widespread use of educa
interventions to eliminate the use of stereo
their success has not been overwhelming.
Increasing the presence of women in traditi
male roles. Having women at the upper levels
organizations can reduce stereotyping. Resear
has shown that the simple presence of a stere
type-disconfirming group member can cause a
change in stereotype-based beliefs (Johnston &
Hewstone, 1990; Weber & Crocker, 1983). Ev
dence in fact suggests that observing successf
women in high-profile, male-typed leadershi
positions not only weakens stereotypical belie
but also activates counterstereotypical belie
(Dasgupta & Asgari, 2004). This outcome is no
necessarilylimitedto upperlevelleadership
positions. Exposure to women in many types o
traditionally male jobs and roles is likely to ch
lenge expectations and erode stereotypical cha
acterizations of women and the resulting lack
fit perceptions that promote gender bias a
discrimination.
However, this outcome is not guaranteed, a
organizations should be aware that merely inc
ing the presence of women is not a foolp
solutionfor changingstereotypes.Sometimes,
for example, women are subtyped and seen
exceptional and not representative of women
general (Richards & Hewstone, 2001), all of w
would not bring about change in how women i
general are viewed. At other times, women

730 Group Processes & Intergroup Relations
seen to have obtained these positions not because
of their merit and qualifications but rather because
of preferentialtreatment(Heilman& Blader,
2001), a situation also unlikely to challenge stereo-
types about women in general. Despite these con-
cerns, however, exposure to successful women
doing jobs that traditionally are done by men pro-
vides information that is at odds with expecta-
tions and, at the very least, provokes a reassessment
of stereotypic beliefs.
Challengingsocial roles. A long-term
approach to tackling gender stereotypes involves
considering their origins. Eagly (1987) theorizes
that gender stereotypes originate from sex roles
that have existed over time. Although a com-
plete overhaul of social roles is outside the pur-
view of organizations, there are some things
they can do to chip away at the thinking that
derivesfrom them. For instance,providing
parental leave for fathers as well as mothers can
contribute to the dismantling of the notion that
women are the parents most suited to take care
of children(Slaughter,2015).Furthermore,
making accessible other work–family options,
including flextime and remote work arrange-
ments, to men and women alike may start to
challenge the idea that women are the caregivers
and men the breadwinners. These roles have
evolved over a long time and are unlikely to dis-
sipate quickly, but organizations can contribute
to their further evolution through the policies
and practices they implement.
Reducing the salience of gender. Even if stereo-
typic beliefs remain intact, reducing the salience
of gender can curtail their use in defining what
a given woman is like and resulting perceptions
of her lack of fit. Some factors that affect gender
salience cannot be feasibly altered, such as moth-
erhood (Heilman & Okimoto, 2008) or physical
attractiveness (Heilman & Stopeck, 1985). How-
ever, decisions involving structure, strategy, and
human resource practices can have an impact on
whether or not gender is salient in a particular
situation and provide avenues for organizations
to limit its activation.
Attending to group composition. In the ques
for diversity, it is important that organizatio
consider not only the numbers of women, b
also their distribution throughout the organi
tion. It is important not to create tokens. In her
pioneering book, Kanter (1977) pointed out tha
numerical scarcity in a group makes the to
woman more distinctive and her gender more s
ent (Cota & Dion, 1986; King, Hebl, George, &
Matusik, 2010). Consistent with this, scarcit
representation has been shown to exacerbate th
mismatch between a woman’s perceived attribu
and the perceived requirements of a male-type
position (Sackett, Dubois, & Noe, 1991). In fact
there is evidence that the same woman is m
likely to be described in stereotypic terms
seen as incompetent when she is in a cohort wi
no other women than when she is in a cohort w
several or many other women (Heilman & Blad
2001). Thus, organizational initiatives that incr
the total number of women in the workplace ca
be undermined if they are not executed in a wa
that takes into account the tokenism problem a
the potential activation of gender stereotyp
produces.
To this end, the practice of clustering wome
can be more effective in avoiding stereotype
activationthan spreadingwomen sparsely
throughout the organization. If implemented
correctly, clustering can undercut the activatio
of stereotypesbecauseviewingwomenuni-
formly in stereotypic terms becomes almost
impossible when there is palpable evidence of
differences among them. This issue also is rele-
vant to the composition of groups and teams.
An organizationwith a limitednumberof
womenmay be inclinedto dispersefemale
employees as much as possible, but this co
inadvertently create a situation in which ma
differentgroupsin the organizationcontain
female tokens. But not all efforts at cluster
will have the desired consequences. The organi
zational area in which the clustering occurs is o
critical importance in determining the effective
ness of a clustering strategy. If clustering serv
only to further group women into stereotyp
cally female areas, such as human resource

Heilman and Caleo 731
marketing,or publicrelations,and not into
those that typically are dominated by men, it will
do little to undercut the activation of stereo-
types and could actually exacerbate it.
Keeping blind. Researchers have found that
the mere provision of a male or female name can
induce the use of gender stereotypes and gender
bias through lack of fit processes (e.g., Heilman
& Haynes, 2005). If a name alone is sufficient to
activate gender stereotypes, one possibility is to
eliminate identifiers from applications. Leaving a
person’s name or identification out of materials
precludes decision makers from making gender
stereotypical inferences because they lack knowl-
edge of the person’s gender. One notable exam-
ple by Goldin and Rouse (2000) involved blind
auditions for an orchestra, where a screen was
used to conceal the identity of the candidate.
The researchers found blind auditions to result
in a higher number of women selected for the
position.
Though it would be unfeasible for organiza-
tions to hide an applicant’s identity during an
interview, it is certainly possible to eliminate iden-
tifying information during the early stages of the
selectionprocess—applications,résumés,and
cover letters. To the extent that it is possible, gen-
der identifiers also can be eliminated from reports
and other work outcomes until evaluations have
been made. However, these procedures should be
engaged in with care. Removing gender markers
can increase the propensity for people to search
for cues that may be related to gender, making
them less likely, for instance, to excuse an appli-
cant with unexplained time off from work in a
résumé. Although more research is needed to
identify when this is most likely to happen, exist-
ing work hints that blind résumés may not always
result in more minority hires (Behaghel, Crepon,
& Le Barbachon, 2015).
Highlighting other ingroup affiliations. People
have many group affiliations—gender is just one
of them. Because gender is so noticeable and
information about it so readily available, it is very
often a primary dimension along which people
are categorized (Ito & Urland, 2003; Stang
Lynch, Duan, & Glas, 1992). But in many situa
tions, cross-cutting ingroup affiliations, such
school attended or place of birth, can superse
gender (van Bavel & Cunningham, 2009). T
malleability of group identity has been am
demonstrated by social identity theorists, a
Tajfel’s (1970) minimal group paradigm ma
clear the feasibility of creating alternative ing
affiliations. Thus, it should be possible to distr
from the salience of gender by emphasizing e
ments of identity other than gender, such
organizational function (e.g., accountant), orga
zational unit (e.g., sales), or organizational loc
(e.g., the uptown group), when evaluative
sion making is occurring. This can be achieved
the type of information about a woman that is
made available to decision makers and the pro
nence and priority given to different aspects o
her identity. Yet, organizations must exercise
tion to ensurethat the emphasizedidentity
dimension is not itself gendered. Some ingrou
affiliations, such as engineer or parent (Heilm
& Okimoto, 2008), can inadvertently increase
salience of gender, therefore working against
intended effect.
Avoiding overt emphasis on diversity goals a
policies. Ironically, despite the positive con
quences of diversity goals and policies (Thom
2004), the use of the diversity label can activa
stereotypes by making salient a woman’s gend
In a study focusing on this issue, a female grou
memberwas characterizedas lesscompetent
when she participated in a workgroup that
assembled to heighten the organization’s dem
graphic diversity as opposed to workgroups th
were assembled either to ensure the organ
tion’s best resources and expertise were re
sented or on the basis of convenience (Heilma
& Welle, 2006). These findings make clear tha
focus on diversity to the exclusion of otherfac
tors can make gender salient, activating st
typesand negativelyaffectingthe ostensible
beneficiary of the diversity effort. A similar ca
can be made for affirmative action program
which also highlight gender and activate gend

732 Group Processes & Intergroup Relations
stereotypes and can promote a “stigma of incom-
petence” for their intended beneficiaries (Heil-
man, Block, & Stathatos, 1997). This also has
implications for quota policies, now so prevalent
internationally. Sometimes an explicit focus on
diversity can backfire, with unintended but none-
theless detrimental, consequences for stereotype
activation.
Changing Perceptions of the Field or Job as Male Gender-
Typed. The second component of the lack of fit
model is the perception of the job. Not only are
some positions populated primarily by men, but
they also are perceived to require stereotypically
male attributes and behaviors that are consistent
with what men are thought to be like and incon-
sistent with what women are thought to be like
(Cejka & Eagly, 1999; Smyth & Nosek, 2015).
There is much evidence, for instance, that the
qualities that are thought necessary to be a suc-
cessful manager overlap with aspects of the male
stereotype but not with aspects of the female ste-
reotype (V. E. Schein, 2001). Consistent with lack
of fit ideas, there is a wealth of research showing
that gender discrimination occurs more frequently
in jobs that are seen as male gender-typed (Koch
et al., 2015; Lyness & Heilman, 2006; Pazy &
Oron, 2001). Although there is some evidence
that there have been shifts in gender-typing, such
as in the lessening of the masculine construal of
the leadership role (Koenig, Eagly, Mitchell, &
Ristikari, 2011), there remains a well delineated set
of jobs and positions that are considered to be
“male” and, more often than not, they are the
ones that are higher paying and more prestigious.
Because stereotypes are most problematic to
women in jobs that are considered male gender-
typed, organizations can work to alter the percep-
tion of the particular position or occupation.
Althoughit may be unrealisticto expectan
organization to change the total perception of a
position or occupation, it can target a range of
practices that support and reinforce the percep-
tion of a field as male gender-typed. Some of
theseinvolvedecisionstraditionallymadeby
human resource departments and others involve
more pervasive images that are perpetuated by
corporate strategies or institutional contexts
this section, we discuss how organizations c
work to transformgenderedviewsof jobs,
organizations, and industries.
Gender-neutralizingjob descriptionsand job
titles. Decades ago, Bem and Bem (1973) d
cussed the prevalence of gender-biased wordin
in organizations—a practice they found to d
courage men and women from applying to cer-
tain positions. The use of gendered language in
their studies was anything but subtle: job adver
tisements in an actual newspaper were segrega
by sex, and job advertisements for a major
ephone company explicitly sought men for t
position of lineman and women for the position
of operator. Although these practices are m
understated today, research suggests that orga
zations continue to use gender-biased wordi
in job advertisements. Gaucher et al. (2011)
for example, found that job advertisements
male-typed positions used more masculine word
ing (e.g., competitive, dominant) than those
female-typed positions. These discrepancies als
appear in languages that are grammatically
linguistically gendered in nature, with research
in Europe finding that many organizations tend
to convey job titles by using the generic mascu-
line form instead of a gender-neutral equivalen
(Sczesny, Formanowicz, & Moser, 2016).
The implications of these subtle practices ar
striking, contributing to lack of fit percepti
and giving rise to discriminatory decisions a
early as recruitment. Importantly, scrutiny o
actual job responsibilities, and what it takes to
fill them, most often suggests a more gende
balanced picture. Even jobs considered to b
unequivocally male in character, such as firefig
ers, have been found to require behavior that is
considered to be stereotypically female—in this
case, compassion (Danbold & Bendersky, 2015)
Job descriptions that are dominated by male-ge
dered language perpetuate the image that t
positions necessitate stereotypically male beha
ior exclusively—an image that conveys the job a
inappropriate for women. In fact, research indi-
cates that gendered language in job description

Heilman and Caleo 733
affects both women’swillingnessto apply
(Gaucher et al., 2011) and their likelihood of
being selected (Horvath & Sczesny, 2015) for
positions depicted as highly male gender-typed.
In line with these ideas, the United States Marine
Corps—a traditionally male-dominated organiza-
tion—recently modified job titles that contained
the word “man,” opting instead for the gender-
neutral “Marine” (Seck, 2016). This action was
directed toward both increasing the appeal of
these jobs to women and decreasing the negativ-
ity they might encounter when pursuing them.
De-masculinizing organizational culture. Organiza-
tional culture refers to the system of shared val-
ues, assumptions, and norms that drive behavior
in organizations (E. H. Schein, 2010). In effect, it
conveys the values and conduct that are deemed
appropriate in a given organization. Culture also
gives decision makers an idea of the kinds of
employees that may fit in best (Schneider, 1987).
It has been argued that an organization’s cul-
ture can be tinged with masculine or feminine
expectations, independent of discipline or type
of work (Guy & Duerst-Lahti, 1992). If this is
the case, we expect that women may be inadvert-
ently shut out of organizations that have a mas-
culine culture. Cultures that demand a high level
of competition, aggression, or risk-taking, for
example, may be seen as incongruent with the
perception that women lack agency.
The challenge is that culture is difficult to
change (E. H. Schein, 2010). However, being vigi-
lant of the norms in place and the extent to which
they perpetuate a gendered culture could be a
fruitful avenue for organizations to pursue. One
way to start would be to survey the organization’s
core valuesand strategy.Anotherpossibility
involves focusing on the norms that thrive in the
organization and considering whether they neces-
sitate masculinity. For example, something as
simple as the style of meetings may perpetuate
the notion that women do not belong in these
settings. Also included in the conceptualization
of organizational culture are artifacts—physical
symbols, rituals, and stories that convey the cul-
ture of the organization (E. H. Schein, 2010),
such as the bull representing Wall Street,
cigars passed around to celebrate the birth of
child, or client meetings at gentlemen’s clu
Environmentssignalwho belongsin them
(Cheryan, Plaut, Davies, & Steele, 2009), a
organizations should be mindful to ensure tha
excess of masculinity is not being conveyed.
Increasing the presence of women in traditi
roles. We previously proposed that increasing
presence of women in male gender-typed
could help combat stereotypical beliefs. Howe
because gender-typing is in part a result of nu
ical gender proportion (Cejka & Eagly, 199
Smyth & Nosek, 2015), increasing the pres
of women in male-dominated fields can also h
to shift job perceptions. Showcasing women w
have defied expectations by succeeding in
considered to be male in gender-type can
the conception of the job itself, raising questio
about whether it really is so male-typed in
requirements if a woman can do it so well. Ho
ever, as with the changing of stereotypes,
efforts can fall short when the stereotype-defy
woman is subtyped and not seen as repres
tive of women in general and when the proces
obtaining the position is seen as based on pref
ence rather than merit. In these situations, gre
numbers of women in stereotype-disconfirm
jobs would be less likely to challenge the perc
tion of a job as male gender-typed and to brin
about a change in how the job is viewed. None
theless, the presence of women in these posit
captures attention and is apt to at least raise q
tions about the gendered nature of the job.
Precluding Negative Expectations From
Affecting Evaluative Decisions
Thus far we have discussed ways in which org
zationscan combatgenderdiscriminationby
minimizing lack of fit perceptions—strategi
that could prevent the formation of negative p
formance expectations that lead to gender
and, in turn, discriminatory behavior. Howev
given the persistence of both stereotypes abo
women and ingrained beliefs about male-ty

734 Group Processes & Intergroup Relations
work, organizations must prepare themselves for
the possibility that, despite best efforts, stereo-
type-based expectations of women’s incompe-
tence to successfully perform traditionally male
positions may prevail. However, with appropriate
intervention,organizationscan preventthese
bias-producingexpectationsfrom influencing
evaluative decisions.
Eliminating Ambiguity. As complex entities, organ-
izations are filled with ambiguities (March &
Olsen, 1985), and such conditions tend to com-
plicate decision making. From the perspective of
personneldecisions,ambiguityallowsbias to
thrive by encouraging evaluators to “fill in the
missing pieces” with their performance expecta-
tions. It is under these conditions that the pro-
cessing of information is most likely to become
distorted in alignment with expectations (Heil-
man & Haynes, 2008). Evidence supports this
idea, with women being evaluated more nega-
tively than men in male-typed fields when ambi-
guitiesexist in the quantityand qualityof
information (Heilman, Wallen, Fuchs, & Tam-
kins, 2004; Koch et al., 2015), performance crite-
ria (Norton et al., 2004), evaluative structure
(Bauer & Baltes, 2002), and teamwork (Caleo &
Heilman, 2014; Heilman & Haynes, 2005).
These findings make clear that ambiguity ena-
bles bias to flourish and suggest that the dimin-
ishing of ambiguity should cause bias to abate.
This implies that organizations should be mind-
ful of the various ways in which ambiguity can
present itself and should focus on reducing it
whenever possible. Although organizations may
be unable to remove all ambiguity, they certainly
can remove some of it, allowing less room for
expectationsto dominatewhen employment
decisions are made.
Clarifyingperformancecriteria. Onetargetfor
organizational reform is the fuzziness of perfor-
mance criteria often used to evaluate employees.
Performance criteria that are vague and abstract
invite bias and cognitive distortion (Fiske & Tay-
lor, 2013). To remedy this, organizations can strive
to develop and use performance criteria that are
as objectiveas possible—measuringnumber
of sales rather than supervisory ratings of sale
performance, dollar earnings rather than gener
productivity, and work attendance rather tha
conscientiousness. Furthermore, when criteria
inherently vague (e.g., team player, charism
resilient), efforts can be made to specify and co
ify the relevant indicators. There should be agr
ment about what these criteria are and precisio
judging whether someone has met them.
Specifying the weighting of performance crit
ies indicate that people have a tendency to rede
fine the meaning of good performance dependi
on who they are evaluating—either overstating
understating the importance of certain criteria
justify their final evaluations (Norton et al., 200
Uhlmann & Cohen, 2005). This research sugges
that if organizations are to stop negative perfo
mance expectations from tainting performan
evaluation, a clear structure must be applie
the evaluationprocess.Organizationsshould
work to design concrete systems for integrating
performance information so that equal weig
and equal criteria are applied across emplo
Such structure gives evaluators no choice but t
assess employees in a standardized fashion, and
precludes them from using their expectations to
skew the importance of any given set of criter
Increasingthe frequencyof performanceevalua-
tions. Researchsuggeststhat memorycan be
swayed by stereotype-based expectations. In
viduals are more likely to recall those pieces of
informationthat alignwith expectationsand
forget those pieces of information that conflict
(Hollingshead & Fraidin, 2003; Martell, 199
We argue that a longer time between evalu
tions increases the likelihood that memory w
be flawed.Furthermore,evaluationsthat are
infrequent are likely to be based on a host
events and episodes, increasing ambiguity a
the opportunity for bias by giving evaluator
a wider range of instances from which to
and choose in line with their expectations when
appraisingemployeeperformance.Therefore,
appraisals occurring on a quarterly or semiann
basis should limit bias to a greater extent
those conducted annually.

Heilman and Caleo 735
Structuringinterviews. Ambiguityis also apt
to manifest itself in the interview process. The
human resource management literature makes
distinctions between structured and unstructured
interviews,with expertsadvisingagainstthe
use of unstructured interview practices (Dana,
Dawes, & Peterson, 2013). Unstructured inter-
views not only are low in validity (Schmidt &
Hunter, 1998), but they also leave the door open
for bias (Levashina, Hartwell, Morgeson, & Cam-
pion, 2014). Yet, decision makers in organizations
still seem eager to rely on intuition when making
screening decisions, rating unstructured interview
techniques more favorably than their less ambigu-
ous counterpart—the structured interview (Dana
et al., 2013; Highhouse, 2008).
From the perspective of avoiding gender bias, the
adoption of structured interview techniques makes
sense for two reasons. Because it calls for interview-
ers to rely on a standardized list of questions, the
trajectory of the interview is less likely to be swayed
by the interviewer’s expectations. Not only are inter-
viewers prone to see what they want to see; they also
can guide the interview in a direction that allows
them to continue seeing the interviewee in line with
initial expectations (Levashina et al., 2014). Second,
structured interviews employ a standardized rating
scale, which provides more objectivity and makes it
less prone to distortion. Still, it is important to note
that structured interviews are not completely imper-
meable to bias (Purkiss, Perrewe, Gillespie, Mayes, &
Ferris, 2006), but adopting them in place of unstruc-
tured interviews lessens the likelihood of expecta-
tionsbiasingfinaldecisions(Bragger,Kutcher,
Morgan, & Firth, 2002).
Structuringteamwork. Organizationsincreas-
ingly rely on teams. Yet, research suggests that
teamworkcreatesambiguityregardingthe
responsibility of its individual members (i.e.,
source ambiguity). Heilman and Haynes (2005)
found source ambiguity to be detrimental to
women in male-typed fields, such that women
were evaluated as less competent and less respon-
sible for a successful outcome than men when a
project was completed in teams. This tendency,
termedattributionalrationalization,has been
replicated in several male-typed contexts (Caleo
& Heilman, 2014; Sarsons, 2017). But with the
increasing use of teams, can organizations re
tically do anything to minimize it? Heilman an
Haynes (2005) found attributional rationalizat
to be decreased when the team task was struc
tured to make clear the contributions of indiv
ual members. Organizations can draw from su
findings. Structuring tasks to reveal individ
contribution, when possible, will limit the imp
of the evaluator’s expectations on assessm
of which team members are most worthy
credit for the collective outcome.
IncreasingMotivation. Evenin the presenceof
ambiguity, organizations can create conditio
that motivate decision makers not to rely on th
stereotype-basedexpectationswhen forming
judgmentsand evaluations.Stereotype-based
expectations allow decision makers to “go
automatic” and not think too much or too hard
about what they are doing (Fiske & Taylor, 20
Increasing the motivation to be accurate ta
evaluators out of this automatic thinking, resu
ing in more systematic information process
where expectations are unlikely to be given fre
reign. This motivation can stem from conce
relating to self-interest or concerns about imp
sion management. Whatever the source, howe
these motivational concerns cause people to b
more deliberative (Harris, 1994), stopping ste
type-based expectations from dominating in de
sion making.
Increasingself-interest. Whenan evaluatoris
interdependent with the person he or she is ev
uating, there is likely to be more motivation to
accurate in assessments (Clark & Wegener, 20
Fiske, 2000). Such interdependence is likel
occur when the evaluated person’s performan
has distinct implications for the evaluator’s we
being, such as when professors’ reputations ar
affected by the jobs their students attain. Und
these circumstances, the influence of self-inte
is likely to overpower that of expectations, an
caution and deliberation are likely. To do well
themselves it is necessary to be highly critical
the other’s strengths and weaknesses, backgro
and experience,and overallpotential—allof

736 Group Processes & Intergroup Relations
which demands careful scrutiny rather than tak-
ing the easy way out and relying on stereotype-
basedexpectations(Neuberg& Fiske,1987).
Thus, organizational procedures that tie the eval-
uator to the evaluatee in ways that directly impact
the evaluator’s rewards, such as having his/her
pay or promotion prospects be formally or infor-
mally contingent on how effectively the evaluatee
performs, should limit the sway of expectations
when judgments are made.
Creating transparency. People tend to want to
“do the right thing,” and they want to be viewed
positively. There is increasing societal pressure
not to be biased in our interactions and, for many,
being anything less than evenhanded in the treat-
ment of others is not acceptable (Dovidio &
Gaertner, 2004; Plant & Devine, 1998). Thus,
there is motivation to demonstrate to others and
to oneself one’s lack of gender bias, and this moti-
vation prompts careful and systematic processing
of information to arrive at an assessment that
will not be embarrassing or potentially tarnish-
ing of one’s image as a fair and decent individual.
Organizations can appeal to these motivations by
introducing more transparency in evaluative deci-
sion-making processes (Castilla, 2015), removing
safeguards that may obscure or hide these pro-
cesses, and making them visible to all. Evaluators’
understanding that what they do will be observ-
able to others is likely to put them on notice and
increase efforts to make a good impression (Mero,
Guidice, & Brownlee, 2007), mitigating against
the use of expectations in making judgments.
Furthermore, salaries themselves could be made
visible to anyone in the organization—a practice
that has been recommended as a way of reducing
the gender wage gap (Kim, 2015).
Requiringaccountability. Whenthey have to
provide an explanation for a decision they make,
peoplebecomemore thoughtfulabout the
actions they take (Scholten, van Knippenberg,
Nijstad, & De Dreu, 2007). Asking decision
makers to justify their decisions to supervisors,
peers, subordinates, or upper management fuels
the motivation to be accurate so that they appear
not only fair-minded but also competent and
reliable (Mero et al., 2007). Thus, accountab
ity not only inhibits the use of expectations in
evaluations, but it also induces the use of more
complex and systematic decision-making stra
egies (Lerner & Tetlock, 1999; Tetlock, 198
Research has indeed demonstrated that peo
who are held accountable make more of an
effort to seek out information, are more atten-
tive when observing performance, and take mo
extensive notes when gathering information, al
of which increases their accuracy (Mero, Moto
idlo, & Anna, 2003). This suggests that account
ability can curb the effect of stereotype-bas
expectations on evaluative judgments. Recen
evidence supports the idea that accountabili
lessens gender bias (Koch et al., 2015). In a lar
scale field study, Castilla (2015) found the intro
duction of accountability into an organization’s
performance–reward system to result in few
pay inequities among women and minorities.
There are several things organizations can do
to create accountability. Organizations must fir
ensurethat decisionmakersare ultimately
accountableto someone—whetherit is the
employee they are rating, a supervisor, or a com
mittee that monitors decisions (Castilla, 201
Furthermore, when evaluators engage in hiring
promotion, or compensation decisions, organiza
tions can require them to report the reason
underlying their evaluations and the decisio
they make. However, reporting alone does n
suffice;becauseindividualstend to shift the
weights they apply to performance criteria
rationalize their decisions (Norton et al., 2004),
is imperative that systems for accountability
set up alongside well-defined and unambigu
systems of performance evaluation. Moreove
decision makers need not only be motivated
make the right decision, but they also must
able to do so. Research suggests that cogni
busyness increases decision makers’ reliance
stereotypes (Gilbert & Hixon, 1991), even unde
mining the positive effects of motivational inte
ventions(Pendry& Macrae,1994).Because
requiring accountability potentially increases th
cognitive demands placed upon decision maker