in press, Policy Insights from the Behavioral and Brain Sciences 1
Implicit Bias and Anti-Discrimination Policy
University of Texas at Austin
University of California, Davis
University of California, Davis
The science behind implicit bias tests (e.g., Implicit Association Test) has become the target of increased criticism.
However, policy-makers seeking to combat discrimination care about reducing bias in people’s actual behaviors, not about
changing a person’s score on an implicit bias test. In line with this argument, we postulate that scientific controversies
about implicit bias tests are irrelevant for anti-discrimination policy, which should instead focus on implicit bias in actual
discriminatory behavior that occurs outside of awareness (in addition to instances of explicit bias). Two well-documented
mechanisms can lead to implicit bias in actual discriminatory behavior: biased weighting and biased interpretation of
information about members of particular social groups. The policy relevance of the two mechanisms is illustrated with their
impact on hiring and promotion decisions, jury selection, and policing. Implications for education and bias intervention are
Keywords: bias intervention; biased information processing; discrimination; implicit bias; policy
On April 12, 2018, two African American men
asked to use the restroom at a Starbucks in
Philadelphia. A barista told them that the bathrooms
were for customers only (Park, 2018). When the two
men were asked to leave the premises after they
occupied a table without making a purchase, they
declined to leave, saying they were waiting for an
acquaintance. In response, the store manager called
the police, who escorted the two men out of the
coffee shop. When a video of the incident taken by a
customer went viral on social media, Starbucks
apologized and closed all of its brand-operated stores
for half a day to provide mandatory implicit bias
training for its 175,000 employees (Chapell, 2018).
In line with Starbucks’s response to the
described incident, an increasing number of public
institutions and private corporations offer or require
implicit bias training for their employees. Yet, at the
same time, the science behind implicit bias tests has
become the target of increased criticism. This
criticism is based on research suggesting that (1)
relations between people’s responses on implicit bias
tests (e.g., Implicit Association Test; Greenwald,
McGhee, & Schwartz, 1998) and actual
discriminatory behavior (e.g., biased hiring
decisions) are rather weak (Oswald, Mitchell,
Blanton, Jaccard, & Tetlock, 2013), (2) many lab-
based interventions influence responses on implicit
bias tests without affecting discriminatory behavior
(Forscher, Lai, Axt, Ebersole, Herman, Devine, &
Nosek, 2019), and (3) responses on implicit bias tests
may reflect the level of bias in a person’s social
environment rather than personal animosities (Payne,
Vuletich, & Lundberg, 2017). Over the past few
years, these concerns have also received increased
attention in the popular media, which is reflected in
critical headlines such as Can We Really Measure
Implicit Bias? Maybe Not (Bartlett, 2017) or The
False ‘Science’ of Implicit Bias (MacDonald, 2017).
Although some of the arguments against implicit
bias tests can be criticized for ignoring important
theoretical, empirical, and methodological issues (see
Brownstein, Madva, & Gawronski, in press;
Gawronski, 2019; Kurdi et al., 2019), the ongoing
controversies surrounding these tests raise the
question of whether it is wise to base anti-
discrimination policies on the notion of implicit bias.
In the current article, we argue that criticism of
implicit bias tests have implications for anti-
discrimination policy only if implicit bias is equated
with responses on these tests (e.g., when implicit bias
is equated with people’s responses on the Implicit
Association Test). Although this conceptualization is
widespread in the scientific literature, it is
problematic for various reasons (see Calanchini &
Sherman, 2013; Corneille & Hütter, in press; De
Houwer, 2019; Gawronski, 2019; Payne & Correll, in
An alternative conceptualization that seems
superior for applied questions (i.e., policy) defines
implicit bias in terms of actual discriminatory
behavior. According to this conceptualization,
discriminatory behavior represents an instance of
implicit bias to the extent that the person showing the
behavior is unaware that their behavior is biased. The
central argument is that anti-discrimination policy
should consider evidence for implicit bias in terms of
this behavioral conceptualization instead of
dismissing the notion of implicit bias because of
extant controversies about implicit bias tests.
in press, Policy Insights from the Behavioral and Brain Sciences 2
Sources of Implicit Bias in Behavior
From a psychological perspective, discrimination
can be said to occur when a person’s behavior toward
a target individual is influenced by the target’s group
membership, including (but not limited to) the
target’s race, gender, or sexual orientation. Of
particular concern for policy are instances of
discrimination that involve negative outcomes for the
Examples include discrimination
based on race or gender in hiring, recruitment,
compensation, promotion, and termination; other
examples include discrimination in housing and
police support (Greenwald & Pettigrew, 2014). In
terms of the above conceptualization, discriminatory
behaviors in these cases are instances of implicit bias
to the extent that a person is unaware that their
behavior is influenced by the category membership of
the target (e.g., the target’s race or gender; see
Gawronski & Bodenhausen, 2012). Research in
social psychology has documented two mechanisms
that can lead to implicit bias in terms of the proposed
conceptualization: (1) biased weighting of mixed
information and (2) biased interpretation of
ambiguous information. The general pattern
underlying both instances is that people show an
initial response to the target that is influenced by the
target’s category membership, and this initial
response influences the subsequent processing of
information about the target.
One mechanism that can lead to implicit bias in
terms of the proposed behavioral conceptualization is
biased weighting of information (e.g., Hodson,
Dovidio, & Gaertner, 2002; Norton, Vandello, &
Darley, 2004; Uhlmann & Cohen, 2005). Such biases
tend to be particularly pronounced in cases involving
judgments and decisions about multiple targets when
the available information about these targets is
mixed. For example, in hiring decisions involving a
male and a female candidate with distinct job-
relevant qualifications, an interviewer may attribute
greater weight to the unique strengths of the male
candidate (e.g., better grades) compared to the unique
strengths of the female candidate (e.g., more
experience). However, the differential weighting of
strengths might be biased in the sense that it merely
Note that a purely psychological definition of discrimination does
not cover systemic aspects (e.g., the lingering consequences of
slavery, redlining, and the denial of civil rights), which we deem
equally important for policy, yet are beyond the scope of this
article. Although the current article focuses mainly on the
psychological level, we deem policies that treat everyone equal
regardless of group membership as insufficient, because such
policies tend to perpetuate existing inequalities rooted in systemic
discrimination (see Rothstein, 2017).
serves as a post-hoc justification for hiring the male
candidate rather than as an a priori criterion. For
example, an interviewer might have an “intuitive”
preference for a male over a female candidate,
because there is a greater fit between social
stereotypes about men and the qualities believed to
be necessary for successful performance (Heilman,
2012). In such cases, the interviewer might
rationalize their “intuitive” preference for the male
candidate by focusing on unique strengths of the
male candidate and/or unique weaknesses of the
female candidate. To the extent that people are
unaware of their bias in weighting mixed information
in a manner that merely justifies a pre-existing
preference, it can lead to discriminatory behavior in
terms of the proposed conceptualization of implicit
Empirical evidence for biased weighting of
mixed information comes from a number of decision-
making studies in which (1) participants were
presented with sets of distinct information about two
(or more) target individuals who differ in terms of
their category membership (e.g., race, gender), and
(2) the assignment of the information sets to the two
targets was experimentally manipulated, such that
participants in one condition saw Information X
about Target A and Information Y about Target B,
while participants in the other condition saw
Information Y about Target A and Information X
about Target B. A key aspect of these studies is that
the two sets of information suggest distinct qualities
in the sense that one set suggests a unique strength in
one domain whereas the other set suggests a unique
strength in a different domain. A biasing effect of the
target’s category membership on participants’
relative weighting of these strengths can be inferred
when participants (1) show a preference for the same
target regardless of the information paired with the
target (e.g., a preference for a male over a female
candidate regardless of the information about the two
candidates) and (2) justify their preference with the
unique strength that happens to characterize the
preferred candidate in the experimental condition
For example, in a study by Norton et al. (2004),
participants viewed application materials of a male
and a female job candidate and indicated which of the
two candidates they would prefer for particular job.
In one condition, the male candidate had less work
experience but more education than the female
candidate did. In another condition, the male
candidate had less education but more work
experience than the female candidate did. Consistent
with the idea of biased weighting, participants
showed a preference for the male candidate in both
experimental conditions and justified their responses
in press, Policy Insights from the Behavioral and Brain Sciences 3
with whatever qualification made him superior to the
female candidate. That is, when the male candidate
excelled in terms of education, participants listed
education as the most significant criterion. Yet, when
the male candidate excelled in terms experience,
participants listed experience as the most significant
criterion (for similar findings, see Hodson et al.,
2002; Uhlmann & Cohen, 2005). Further research
suggests that biasing effects of differential weighting
occur outside of awareness, in that participants’ self-
perceptions of objectivity in their decision were
associated with greater (rather than smaller) bias
(Uhlmann & Cohen, 2005).
Although biased weighting can lead to
discrimination in a wide range of real-world contexts,
its effects are most prominently reflected in selective
choice decisions, such as admission, hiring, and
promotion decisions. In such cases, decision-makers
often have to identify a small number of candidates
(or only one) among a large number of highly
qualified candidates. What makes these decisions
particularly difficult is that the relevant evaluation
criteria are often multidimensional rather
unidimensional, forcing decision-makers to compare
“apples and oranges” when candidates differ in term
of their relative strengths. Thus, to the extent that the
relative importance of evaluation criteria remains
unspecified, decision-makers have to come up with
their own weighting schema, leaving considerable
room for arbitrary weightings that merely justify a
decision-maker’s biased preference (Bragger,
Kutcher, Morgan, & Firth, 2002; Uhlmann & Cohen,
2005). Such biases are difficult to address, because
decision-makers tend to think of their decisions as
being based on their impressions of specific
individuals rather than beliefs about the social groups
to which these individuals belong (see Ledgerwood,
Eastwick, & Gawronski, in press). For example,
people may deny that gender had any influence on
their preference for a male over a female candidate
and refer primarily to unique strengths of the male
target without realizing that they would justify their
preference with whatever criterion makes the male
candidate seem superior.
Another example of biased weighting in real-
world contexts is bias in jury selection. In 1986, the
U.S. Supreme Court ruled that prospective jurors
could not be challenged on the basis of being a
member of a cognizable racial group (Batson v.
Kentucky, 1986). Subsequent rulings have extended
this rule to preemptory challenges based on gender
(J.E.B. v. Alabama, 1994). However, questions have
been raised about whether requiring attorneys to
justify suspicious challenges—which has become
common practice since Batson v. Kentucky—is
effective in preventing bias in jury selection
(Sommers & Norton, 2008). Similar to the concern
about biased weighting in the justification of hiring
decisions, attorneys may justify their preemptory
challenges by referring to race- and gender-neutral
characteristics, but this does not mean that their
challenges are unaffected by a juror’s race and
gender. In line with this concern, experimental
studies found that race influenced preemptory
challenges by advanced law students and practicing
attorneys, but their justifications were entirely race-
neutral (Sommers & Norton, 2007). Although
participants might have been aware of their biased
reasoning, biased weighting of information to justify
a particular decision would qualify as an instance of
implicit bias, to the extent that attorneys are unaware
of the influence of race or gender on their preemptory
Even when two individuals do the same thing,
people often perceive the behavior differently
depending on the category membership of the
behaving person (e.g., Darley & Gross, 1983;
Duncan, 1976; Gawronski, Geschke, & Banse, 2003;
Hugenberg & Bodenhausen, 2003; Kunda &
Sherman-Williams, 1993; Sagar & Schofield, 1980;
Trope, 1986). Such biased perceptions are
particularly pronounced when the observed behavior
is ambiguous. For example, a teacher may perceive a
student’s essay for an English class as stronger when
the student is White than when the student is Black,
but the student’s race may have little impact on the
teacher’s perceptions of objectively correct or
incorrect responses on a math exam (Darley & Gross,
1983). Because people tend to treat their subjective
perceptions as direct reflections of objective reality
rather than the product of active construal processes
that are prone to perceptual biases (Trope & Gaunt,
1999), attempts to correct one’s biased perceptions
are relatively rare, leading to discriminatory behavior
without people being aware of their biases (see
Strack & Hannover, 1996; Wegener & Petty, 1997;
Wilson & Brekke, 1994).
Empirical evidence for biased interpretations
comes from a number of studies in which (1)
participants were presented with ambiguous
information about a target person and (2) the target
person’s category membership was experimentally
manipulated, such that the target belonged to one
social category (e.g., White) in one condition and a
different social category (e.g., Black) in another
condition. A key aspect of these studies is that the
ambiguous information is exactly the same in the two
experimental conditions, the only difference being
the category membership of the target. A biasing
effect of the target’s category membership on
participants’ interpretations of the ambiguous
in press, Policy Insights from the Behavioral and Brain Sciences 4
behavior can be inferred when participants judge the
behavior differently in the two experimental
For example, in a study by Hugenberg and
Bodenhausen (2003), participants watched short
video clips of either Black or White targets whose
facial expressions changed either from smiling to
frowning or from frowning to smiling. The
experimenters created the target faces with a 3D
computer program, such that the facial structure was
identical for matched Black and White targets, the
only difference being their skin color and hairstyle.
Participants’ task was to press a key (1) as soon as
they saw hostility in the target’s face, when the facial
expression was changing from smiling to frowning,
and (2) as soon as they do not see any hostility in the
target’s face, when the facial expression was
changing from frowning to smiling. Consistent with
the hypothesis that even perceptions of basic
emotional expressions can be biased by category
membership, participants perceived hostility earlier
and for longer durations when the target faces were
Black than when they were White (see also Bijlstra,
Holland, Dotsch, Hugenberg, & Wigboldus, 2014;
Hutchings & Haddock, 2008). Further research
suggests that such biasing effects occur outside of
awareness, in that even people who are highly
motivated to respond in a non-prejudicial manner
show the same bias in their perceptions of ambiguous
information (Gawronski et al., 2003).
The real-world relevance of biased
interpretations can be illustrated with the cases listed
under hashtag #LivingWhileBlack, which describe
ordinary activities for which police have been called
on African Americans (Griggs, 2018). In addition to
the above-mentioned case of waiting for an
acquaintance at Starbucks, the list includes mundane
activities such as moving into an apartment, making a
phone call in a hotel lobby, shopping for prom
clothes, not waving while leaving an Airbnb, eating
lunch on a college campus, working as a home
inspector, and delivering newspapers. The general
theme underlying these cases is that, while the
described behaviors tend to be perceived as ordinary
when a White person does them, they are perceived
as suspicious (and potentially threatening) when a
Black person does them.
A lethal variant of such biased perceptions is the
tendency to more frequently misidentify harmless
objects as weapons when they are held by a Black
person than when they are held by a White person
(for a review, see Payne & Correll, in press).
Although early research suggested that this tendency
is rooted in impulsive response tendencies that can be
intentionally controlled given sufficient time and
mental resources (Payne, Shimizu, & Jacoby, 2005),
more recent evidence supports the idea that the
greater tendency to shoot unarmed Black (vs. White)
men is at least partly driven by unconscious visual
processes leading to biased perceptions of ambiguous
objects (Correll, Wittenbrink, Crawford, & Sadler,
2015). Beyond racially biased identifications of
harmless objects as weapons, unconscious perceptual
biases have also been implicated in divergent
perceptions of video evidence (Granot, Balcetis,
Feigeson, & Tyler, 2016).
Another illustrative example is the concern that
the same agentic behavior is often perceived less
favorably when a woman does it than when a man
does (Rudman, Moss-Racusin, Glick, & Phelan,
2012). For example, while self-promoting, assertive,
and dominant behavior is often interpreted positively
in a man (e.g., reflecting confidence and leadership),
the same behavior is more likely to be interpreted
negatively in a woman (e.g., reflecting neuroticism
and disagreeableness). In work contexts, such biased
perceptions can lead to gender discrimination in
promotions for leadership roles, given that promotion
decisions depend on perceptions of leadership-
relevant traits. Yet, unlike the idea that gender
influences such decisions in a direct manner, the
notion of biased interpretation suggests a more
subtle, indirect effect. That is, a person’s gender
influences people’s perceptions of the person’s
behavior, which in turn influences overall
impressions of that person’s suitability for a
leadership role (Trope, 1986). As with the effects of
biased weighting, such biases are difficult to address,
because decision-makers tend to think of their
decisions as being based on their impressions of a
specific person rather than their beliefs about men
and women in general (see Ledgerwood et al., in
press). Thus, people may deny that a target’s
category membership had any influence on their
decision and refer primarily to their perceptions of
the specific target person, without realizing that their
perception of the target’s behavior is influenced by
the target’s category membership (see Dovidio &
For example, a manager might carefully select a
set of qualities that an employee should display in
order to get a promotion (e.g., assertiveness, strong
leadership potential) and then evaluate each
employee with respect to those traits. Yet, implicit
bias could creep into this decision if the manager
perceives the same behavior differently depending on
the group membership of the employees (e.g., Mark
and Maria both express anger toward someone who
missed a deadline, but Mark’s behavior is interpreted
as assertive whereas Maria’s behavior is interpreted
as volatile; Mark is then evaluated as more assertive
and thus more deserving of a promotion). Thus, even
in press, Policy Insights from the Behavioral and Brain Sciences 5
when people are careful to be evenhanded in their
decision-making process, biased interpretations of
ambiguous behavior may have already shaped
upstream impressions of the individuals being
Implications for Education and Intervention
Organizational efforts to combat bias have
created a multi-billion dollar industry (Lipman,
2018). Yet, empirical assessments of their
effectiveness in increasing diversity suggest a bleak
conclusion (Kalev, Dobbin, & Kelly, 2006).
Although the identified reasons for this outcome are
complex and beyond the scope of this article (for a
discussion, see Carter, Onyeador, & Lewis, in press),
the reviewed effects of biased weighting and biased
interpretation suggest that extant interventions would
benefit from considering their contributions to
discrimination in the workplace and various other
A first step in this regard is to increase public
awareness of the two sources of bias by educating
people how biased weighting and biased
interpretation can lead to discriminatory behavior.
Examples of suitable contexts for this endeavor are
organizational trainings and dedicated lectures in
high-school classes, which may include presentations
on the evidence reviewed above. Hands-on exercises
that replicate experimental demonstrations of the two
mechanisms could be particularly helpful to illustrate
their impact. Popular media may also contribute to
increasing public awareness by communicating the
scientific evidence for biased weighting and biased
interpretation to non-academic audiences. Because
describing bias as unconscious can lead people to feel
less accountable for biased actions (Daumeyer,
Onyeador, Brown, & Richeson, 2019; Payne,
Cameron, & Knobe, 2010), discussions of implicit
bias should emphasize the responsibility of
individuals and organizations to create policies and
procedures to prevent expressions of implicit bias in
individual behavior. To avoid implying that bias only
exists at the level of individuals, these discussions
should also contextualize the issue of implicit bias at
the individual level in a broader understanding of
systemic and historical bias (see Bonam, Das,
Coleman, & Salter, 2019; Salter, Adams, & Perez,
Strategies for Individuals
Although knowledge of the two mechanisms that
we have described is an important first step in
combatting their effects, such knowledge alone seems
unlikely to eliminate their impact without additional
hands-on strategies (Carter et al., in press). For
example, a person may be aware that biased
weighting can lead to discrimination in hiring
decisions, but the person may not be aware that
biased weighting influences their own hiring decision
in a particular case. Regarding bias correction at the
individual level, some research suggests that a
strategy termed consider-the-opposite (Hirt &
Markman, 1995; Lord, Lepper, & Preston, 1984) can
be helpful to combat effects of biased weighting. The
strategy involves a reconsideration of the same
information assuming that the target differed on a
potentially biasing characteristic. For example, in
cases involving a choice between a male and a female
job candidate, people may mentally simulate whether
they would make a different choice if the
qualifications of the two candidates were swapped. If
people realize that their preference for the male
candidate would be unaffected by a swap of
qualifications, their formerly “implicit” bias would
become “explicit” in the sense that they are now
aware of the biasing effect of gender on their hiring
preference. This insight allows decision-makers to
“re-compute” their judgments taking the identified
source of bias into account (Strack, 1992).
Although mental simulations considering the
opposite can be helpful in identifying effects of
biased weighting, identifying effects of biased
interpretation is more difficult. For example, in cases
involving interpretations of ambiguous behavior
shown by an African American person, people may
mentally simulate how they would perceive the
behavior if the target was White. To the extent that
the behavior would be perceived differently for a
White target, people would become aware of the
biasing effect of race on their perception of the
target’s behavior, providing a basis to “re-compute”
their judgments taking the identified source of bias
into account (Strack, 1992). However, the likelihood
of such awareness-raising effects is relatively low,
because such mental simulations are based, not on
“objective” features of the observed behavior, but
subjective interpretations of the behavior, which are
prone to the bias described above. For example, a
person may conclude that they should call the police
on anyone who is trying to break into a house
regardless of whether person is White or Black.
However, they may not realize that they are
interpreting the target’s ambiguous behavior as
“trying to break into a house” only because the target
is Black, and that they would not have interpreted the
Although it is possible that some people respond defensively to
the outcomes of their mental simulations and try to justify their
initial preference, any such justifications will differ from the initial
ones, because people would have to justify a bias that is now
explicit (e.g., they would have to justify why they would hire a
male over a female candidate regardless of their qualifications).
in press, Policy Insights from the Behavioral and Brain Sciences 6
same behavior in this way if the target had been
White. This intricate link makes it difficult to
determine if one’s perception of a person’s behavior
is biased by the person’s category membership.
Strategies for Organizations
As the discussion above makes clear, identifying
and correcting for implicit bias at the individual level
can be challenging. Indeed, a more effective way to
combat implicit bias is to change structures and
procedures to create contexts in which discrimination
is less likely to occur (Carter et al., in press; Salter et
al., 2018). One of the most effective strategies in
decision-making contexts is to “remove” potentially
biasing category information, as is the case in the
practice of blinded evaluation. Such a policy can
effectively prevent effects of both biased weighting
and biased interpretation (Goldin & Rouse, 2000). If
there is no category information to begin with, it
cannot bias the weighting of mixed information or
the interpretation of ambiguous information.
To the extent that blinding is not feasible, an
alternative strategy to prevent effects of biased
weighting is to specify unambiguous decision criteria
before decision-makers review any information about
the relevant target individuals (Bragger et al., 2002;
Uhlmann & Cohen, 2005). In hiring contexts, clear
specifications and prior commitment to specific
criteria can reduce arbitrary weightings that serve to
merely justify a pre-existing preference independent
of the actual information about the candidates
(Uhlmann & Cohen, 2005). Similar effects occur for
highly structured (compared to informal) interviews,
which have proven their effectiveness in reducing
biases against pregnant job applicants (Bragger et al.,
However, prior specification of evaluation
criteria will increase diversity only if the identified
criteria are unbiased in the sense that they do not
favor members of certain groups. For example, a
manager might select a set of qualities for evaluating
employees that includes the traits assertive,
confident, and leadership potential. Such a list can
lead to biased outcomes if the identified qualities are
more readily inferred from behaviors when the
person performing the behavior is a man rather than a
woman (e.g., via biased interpretations of ambiguous
behavior). To combat this source of bias, decision-
makers would need to be accountable for adding
equally desirable qualities that fit better with
stereotypes of women than men (e.g., excellent
communicator and inspires effective teamwork), so
that the resulting list of desired criteria became more
balanced. It may also help to create procedures that
increase the amount of time that evaluation
committees spend discussing attributes that favor
systematically disadvantaged candidates (e.g., asking
committees to spend as much time discussing
candidate warmth as they spend discussing candidate
competence), although additional research is needed
to test this intervention idea in real-world hiring
contexts (Chang & Cikara, 2018).
Of course, the psychological processes
underlying discrimination do not take place in a
vacuum. Individual decisions and behaviors are
always situated in a broader historical and societal
context. Strategies designed to combat implicit bias
at the individual level can only go so far (Bonam et
al., 2019; Payne & Vuletich, 2018). It will be
important for organizations to invest in long-term
training (rather than expecting a single training to
have long-term behavioral consequences), monitor
training effectiveness in particular contexts, and
develop organizational structures that increase
accountability for diversity (e.g., diversity
committees and staff positions; Carter et al., in press;
Kalev et al., 2006). Even perfectly evenhanded
behavior at the individual level can perpetuate
inequalities produced by long periods of systemic
discrimination (see Kendi, 2017; Rothstein, 2017).
Because such processes involve societal factors that
go beyond the psychological mechanisms discussed
in the current article, they require additional
strategies to combat bias at the systemic level (e.g.,
affirmative action policies).
Research on implicit bias has become the target
of increased criticism, raising questions about
whether anti-discrimination policy should be based
on a controversial construct. In response to this
concern, we argued that extant criticism of implicit
bias tests (e.g., Implicit Association Test) affects
anti-discrimination policy only if implicit bias is
equated with responses on these tests; it remains
unaffected if implicit bias is defined behaviorally in
terms of actual discriminatory behavior. This
alternative conceptualization highlighted the role of
two well-understood mechanisms that can lead to
discriminatory behavior outside of awareness: biased
weighting of mixed information and biased
interpretation of ambiguous information. Of course,
either type of bias may be systematically related to
responses on implicit bias tests, which is a question
for basic scientific research (for a review, see
Gawronski, Hofmann, & Wilbur, 2006). However,
this question is entirely irrelevant for anti-
discrimination policy on implicit bias. What matters
for such policy is implicit bias in actual
The social psychological literature offers
valuable insights into the mechanisms underlying
implicit bias in actual discriminatory behavior and
in press, Policy Insights from the Behavioral and Brain Sciences 7
potential strategies to combat their effects. Some of
these strategies have already proven their
effectiveness in reducing bias (e.g., blinded
evaluations, prior specification of evaluation criteria,
structured as opposed to informal interviews); others
were derived from lab-based findings that await
further testing in real-world contexts (e.g., public
knowledge of the two mechanisms, consider-the-
opposite, stereotypically balanced evaluation
criteria). Yet, all of them can be easily included in
extant diversity trainings, and organizational
executives can implement them into their decision-
making procedures with little or no extra costs (e.g.,
blinded evaluations, prior specification of evaluation
criteria, structured as opposed to informal interviews,
stereotypically balanced evaluation criteria).
Although effective interventions will require
approaches that target individual, organizational, and
systematic aspects of discrimination, neither
approach will succeed without considering implicit
bias in discriminatory behavior that occurs outside of
Bartlett, T. (2017, January 5). Can we really measure
implicit bias? Maybe not. The Chronicle of
Higher Education. Retrieved from
Really-Measure-Implicit/238807 (April 9, 2020).
Batson v. Kentucky (1986). 476 U.S. 79.
Bijlstra, G., Holland, R. W., Dotsch, R., Hugenberg,
K., & Wigboldus, D. H. J. (2014). Stereotype
associations and emotion recognition.
Personality and Social Psychology Bulletin, 40,
Bonam, C. M., Nair Das, V., Coleman, B. R., &
Salter, P. (2019). Ignoring history, denying
racism: Mounting evidence for the Marley
hypothesis and epistemologies of ignorance.
Social Psychological and Personality Science,
Bragger, J. D., Kutcher, E., Morgan, J., & Firth, P.
(2002). The effects of the structured interview on
reducing biases against pregnant job applicants.
Sex Roles, 46, 215-226.
Brownstein, M., Madva, A., & Gawronski, B. (in
press). Understanding implicit bias: Putting the
criticism into perspective. Pacific Philosophical
Calanchini, J., & Sherman, J. W. (2013). Implicit
attitudes reflect associative, non-associative, and
non-attitudinal processes. Social and Personality
Psychology Compass, 7, 654-667.
Carter, E., Onyeador, I., & Lewis, N. A., Jr. (in
press). Developing and delivering effective anti-
bias training: Challenges and recommendations.
Behavioral Science and Policy.
Chapell, B. (2018, May 29). Starbucks closes more
than 8,000 stores today for racial bias training.
National Public Radio. Retrieved from
training (April 9, 2020).
Chang, L. W., & Cikara, M. (2018). Social decoys:
Leveraging choice architecture to alter social
preferences. Journal of Personality and Social
Psychology, 115, 206-223.
Corneille, O., & Hütter, M. (in press). Implicit? What
do you mean? A comprehensive review of the
delusive implicitness construct in attitude
research. Personality and Social Psychology
Correll, J., Wittenbrink, B., Crawford, M., & Sadler,
M.S. (2015). Stereotypic vision: How
stereotypes disambiguate complex visual stimuli.
Journal of Personality and Social Psychology,
Darley, J. M., & Gross, P. H. (1983). A hypothesis-
confirming bias in labeling effects. Journal of
Personality and Social Psychology, 44, 20-33.
Daumeyer, N. M., Onyeador, I., Brown, X., &
Richeson, J. A. (2019). Consequences of
attributing discrimination to implicit vs. explicit
bias. Journal of Experimental Social Psychology,
De Houwer, J. (2019). Implicit bias is behavior: A
functional-cognitive perspective on implicit bias.
Perspectives on Psychological Science, 14, 835-
Dovidio, J. F., & Gaertner, S. L. (2004). Aversive
racism. Advances in Experimental Social
Psychology, 36, 1-52.
Duncan, B. L. (1976). Differential perception and
attribution of intergroup violence: Testing the
lower limits of stereotyping of Blacks. Journal of
Personality and Social Psychology, 34, 590-598.
Forscher, P. S., Lai, C. K., Axt, J. R., Ebersole, C. R.,
Herman, M., Devine, P. G., & Nosek, B. A.
(2019). A meta-analysis of procedures to change
implicit measures. Journal of Personality and
Social Psychology, 117, 522-559.
Gawronski, B. (2019). Six lessons for a cogent
science of implicit bias and its criticism.
Perspectives on Psychological Science, 14, 574-
Gawronski, B., & Bodenhausen, G. V. (2012). Self-
insight from a dual-process perspective. In S.
Vazire & T. D. Wilson (Eds.), Handbook of self-
knowledge (pp. 22-38). New York: Guilford
in press, Policy Insights from the Behavioral and Brain Sciences 8
Gawronski, B., Geschke, D., & Banse, R. (2003).
Implicit bias in impression formation:
Associations influence the construal of
individuating information. European Journal of
Social Psychology, 33, 573-589.
Gawronski, B., Hofmann, W., & Wilbur, C. J.
(2006). Are “implicit” attitudes unconscious?
Consciousness and Cognition, 15, 485-499.
Goldin, C., & Rouse, C. (2000). Orchestrating
impartiality: The impact of" blind" auditions on
female musicians. American Economic
Review, 90, 715-741.
Granot, Y., Balcetis, E., Feigenson, N., & Tyler, T.
(2018). In the eyes of the law: Perception versus
reality in appraisals of video evidence.
Psychology, Public Policy, and Law, 24, 93-104.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. K.
L. (1998). Measuring individual differences in
implicit cognition: The Implicit Association
Test. Journal of Personality and Social
Psychology, 74, 1464-1480.
Greenwald, A. G., & Pettigrew, T. F. (2014). With
malice toward none and charity for
some: Ingroup favoritism enables
discrimination. American Psychologist, 69, 669-
Griggs, B. (2018, December 28). Living while black:
Here are all the routine activities for which
police were called on African-Americans this
year. CNN. Retrieved from
Heilman, M. E. (2012). Gender stereotypes and
workplace bias. Research in Organizational
Behavior, 32, 113-135.
Hirt, E. R., & Markman, K. D. (1995). Multiple
explanation: A consider-an-alternative strategy
for debiasing judgments. Journal of Personality
and Social Psychology, 69, 1069-1086.
Hodson, G., Dovidio, J. F., & Gaertner, S. L. (2002).
Processes in racial discrimination: Differential
weighting of conflicting information. Personality
and Social Psychology Bulletin, 28, 460-471.
Hugenberg, K., & Bodenhausen, G. V. (2003).
Facing prejudice: Implicit prejudice and the
perception of facial threat. Psychological
Science, 14, 640-643.
Hutchings, P. B., & Haddock, G. (2008). Looking
black in anger: The role of implicit prejudice in
the categorization and perceived emotional
intensity of racially ambiguous faces. Journal of
Experimental Social Psychology, 44, 1418-1420.
J.E.B. v. Alabama (1994). 511 U.S. 127.
Kalev, A., Dobbin, F., & Kelly, E. (2006). Best
practices or best guesses? Assessing the efficacy
of corporate affirmative action and diversity
policies. American Sociological Review, 71, 589-
Kendi, I. X. (2017). Stamped from the beginning: The
definitive history of racist ideas in America. New
York: Random House.
Kunda, Z., & Sherman-Williams, B. (1993).
Stereotypes and the construal of individuating
information. Personality and Social Psychology
Bulletin, 19, 90-99.
Kurdi, B., Seitchik, A. E., Axt, J. R., Carroll, T. J.,
Karapetyan, A., Kaushik, N., Tomezsko, D.,
Greenwald, A. G., & Banaji, M. R. (2019).
Relationship between the Implicit Association
Test and intergroup behavior: A meta-analysis.
American Psychologist, 74, 569-586.
Ledgerwood, A., Eastwick, P. W., & Gawronski, B.
(in press). Experiences of liking versus ideas
about liking. Behavioral and Brain Sciences.
Lipman, J. (2018, January 25). How diversity training
infuriates men and fails women. Time. Retrieved
infuriates-men-fails-women/ (June, 9, 2020).
Lord, C. G., Lepper, M. R., & Preston, E. (1984).
Consider the opposite: A corrective strategy for
social judgments. Journal of Personality and
Social Psychology, 47, 1231-1243.
MacDonald, H. (2017, October 9). The false
“science” of implicit bias. Wall Street Journal.
of-implicit-bias-1507590908 (April 9, 2020).
Norton, M. I., Vandello, J. A., & Darley, J. M.
(2004). Casuistry and social category bias.
Journal of Personality and Social Psychology,
Oswald, F. L., Mitchell, G., Blanton, H., Jaccard, J.,
& Tetlock, P. E. (2013). Predicting ethnic and
racial discrimination: A meta-analysis of IAT
criterion studies. Journal of Personality and
Social Psychology, 105, 171-192.
Park, M. (2018, April 18). What the Starbucks
incident tells us about implicit bias. CNN.
-bias-philadelphia-starbucks/index.html (April 9,
Payne, B. K., Cameron, C. D., & Knobe, J. (2010).
Do theories of implicit race bias change moral
judgments? Social Justice Research, 23, 272-
Payne, B. K., & Correll, J. (in press). Race, weapons,
and the perception of threat. Advances in
Experimental Social Psychology.
Payne, B. K., Shimizu, Y., & Jacoby, L. L. (2005).
Mental control and visual illusions: Toward
in press, Policy Insights from the Behavioral and Brain Sciences 9
explaining race-biased weapon identifications.
Journal of Experimental Social Psychology, 41,
Payne, B. K., & Vuletich, H. A. (2018). Policy
insights from advances in implicit bias research.
Policy Insights from the Behavioral and Brain
Sciences, 5, 49-56.
Payne, B. K., Vuletich, H. A., & Lundberg, K. B.
(2017). The bias of crowds: How implicit bias
bridges personal and systemic prejudice.
Psychological Inquiry, 28, 233-248.
Rothstein, R. (2017). The color of law: A forgotten
history of how our government segregated
America. New York: Liveright Publishing.
Rudman, L. A., Moss-Racusin, C. A., Glick, P., &
Phelan, J. E. (2012). Reactions to vanguards:
Advances in backlash theory. Advances in
Experimental Social Psychology, 45, 167-227.
Sagar, H. A., & Schofield, J. W. (1980). Racial and
behavioral cues in black and white children's
perceptions of ambiguously aggressive acts.
Journal of Personality and Social Psychology,
Salter, P. S., Adams, G., & Perez, M. J. (2018).
Racism in the structure of everyday worlds: A
cultural-psychological perspective. Current
Directions in Psychological Science, 27, 150-
Sommers, S. R., & Norton, M. I. (2007). Race-based
judgments, race-neutral justifications:
Experimental examination of peremptory use and
the Batson challenge procedure. Law and Human
Behavior, 31, 261-273.
Sommers, S. R., & Norton, M. I. (2008). Race and
jury selection: Psychological perspectives on the
peremptory challenge debate. American
Psychologist, 63, 527-539.
Strack, F. (1992). The different routes to social
judgments: Experiential versus informational
strategies. In L. L. Martin & A. Tesser (Eds.),
The construction of social judgments (pp. 249-
275). Hillsdale, NJ: Erlbaum.
Strack, F., & Hannover, B. (1996). Awareness of the
influence as a precondition for implementing
correctional goals. In P. M. Gollwitzer & J. A.
Bargh (Eds.), The psychology of action: Linking
cognition and motivation to behavior (pp. 579-
596). New York: Guilford Press.
Trope, Y. (1986). Identification and inferential
processes in dispositional attribution.
Psychological Review, 93, 239-257.
Trope, Y. & Gaunt, R. (1999). A dual-process model
of overconfident attributional inferences. In S.
Chaiken & Y. Trope (Eds.), Dual-process
theories in social psychology (pp. 161-178). New
York: Guilford Press.
Uhlmann, E. L., & Cohen, G. L. (2005). Constructed
criteria: Redefining merit to justify
discrimination. Psychological Science, 16, 474-
Wegener, D. T., & Petty, R. E. (1997). The flexible
correction model: The role of naive theories of
bias in bias correction. Advances in Experimental
Social Psychology, 29, 141-208.
Wilson, T. D., & Brekke, N. (1994). Mental
contamination and mental correction: Unwanted
influences on judgments and evaluations.
Psychological Bulletin, 116, 117-142.
Preparation of this article was supported by
National Science Foundation Grant BCS-1941440.
Any opinions, findings, and conclusions or
recommendations expressed in this material are those
of the authors and do not necessarily reflect the views
of the National Science Foundation.