ArticlePDF Available

Employment Discrimination: The Role of Implicit Attitudes, Motivation, and a Climate for Racial Bias

Authors:

Abstract and Figures

This study is an attempt to replicate and extend research on employment discrimination by A. P. Brief and colleagues (A. P. Brief, J. Dietz, R. R. Cohen, S. D. Pugh, & J. B. Vaslow, 2000). More specifically, the authors attempted (a) to constructively replicate the prior finding that an explicit measure of modern racism would interact with a corporate climate for racial bias to predict discrimination in a hiring context and (b) to extend this finding through the measurement of implicit racist attitudes and motivation to control prejudice. Although the authors were unable to replicate the earlier interaction, they did illustrate that implicit racist attitudes interacted with a climate for racial bias to predict discrimination. Further, results partially illustrate that motivation to control prejudice moderates the relationship between explicit and implicit attitudes. Taken together, the findings illustrate the differences between implicit and explicit racial attitudes in predicting discriminatory behavior.
Content may be subject to copyright.
RESEARCH REPORTS
Employment Discrimination: The Role of Implicit Attitudes, Motivation,
and a Climate for Racial Bias
Jonathan C. Ziegert and Paul J. Hanges
University of Maryland
This study is an attempt to replicate and extend research on employment discrimination by A. P. Brief
and colleagues (A. P. Brief, J. Dietz, R. R. Cohen, S. D. Pugh, & J. B. Vaslow, 2000). More specifically,
the authors attempted (a) to constructively replicate the prior finding that an explicit measure of modern
racism would interact with a corporate climate for racial bias to predict discrimination in a hiring context
and (b) to extend this finding through the measurement of implicit racist attitudes and motivation to
control prejudice. Although the authors were unable to replicate the earlier interaction, they did illustrate
that implicit racist attitudes interacted with a climate for racial bias to predict discrimination. Further,
results partially illustrate that motivation to control prejudice moderates the relationship between explicit
and implicit attitudes. Taken together, the findings illustrate the differences between implicit and explicit
racial attitudes in predicting discriminatory behavior.
Keywords: discrimination, implicit attitudes, racism, prejudice, IAT
In recent years, racist attitudes have evolved from being blatant
and hostile in nature to being more subtle and ambivalent (Brief,
Dietz, Cohen, Pugh, & Vaslow, 2000). Indeed, whereas traditional
self-report measures have indicated that there has been a decline in
racist attitudes, discrimination continues in employment decisions
(Maass, Castelli, & Arcuri, 2000). This discrepancy and shift in the
nature of racist attitudes has prompted social scientists to design
new measures that are consistent with the more modern expression
of racism (McConahay, 1986; McConahay, Hardee, & Batts,
1981). These scales attempt to get around self-presentation bias
and identify individuals with negative racial attitudes by using
questions in which the prejudiced response could be attributed to
“racially neutral ideology” (Fazio, Jackson, Dunton, & Williams,
1995).
Recently, some researchers have moved away from these self-
report measures to physiological, response latency, or priming
measures to assess an individual’s level of racist attitudes (Fazio &
Olson, 2003). These more implicit measures are believed to be less
susceptible to self-presentation biases and thus are more successful
at assessing prejudices. Although research has documented that
these implicit measures correlate with other attitudes and predict
microlevel behavior, there is currently little evidence indicating
that such implicit attitudes are useful for predicting more mac-
rolevel behavior, such as discriminatory hiring decisions.
The present study was designed to be a constructive replication
of the Brief et al. (2000) study, which showed that modern racists
act on their prejudices in particular social climates. Lykken (1968)
defined a constructive replication as a study that not only tests the
validity of prior findings but also tests new hypotheses. The
present study extends the Brief et al. (2000) study in several ways.
First, in addition to assessing modern racism, we included a
measure of more traditional racist beliefs (i.e., old-fashioned or
hostile racism) to ascertain whether these more modern racism
assessments were really needed to identify racist individuals. Sec-
ond, in addition to using these two self-report (i.e., explicit)
measures of racism, we included an implicit racial attitudes mea-
sure. Third, we included a measure of motivation to control prej-
udice to test whether it is indeed a self-presentation bias that
accounts for different results obtained by explicit and implicit
measures. Finally, we used a more sensitive measure of racial
discrimination to test our hypotheses by comparing differences in
the ratings of Blacks and Whites with a more sophisticated statis-
tical technique: hierarchical linear modeling (HLM; Bryk & Rau-
denbush, 1992; Kreft & De Leeuw, 1998). Taken together, these
additions allowed us to test new hypotheses that provide a greater
understanding of employment discrimination. In summary, the
present study merged the literatures on racist attitudes and their
measurement, self-presentation bias, and organizational social
norms and climates to understand the differential utility of implicit
and explicit measures in predicting individuals’ behavior in a
selection context.
Employment Discrimination and Racist Attitudes
Many studies have looked at the relationship between race
considerations and employment discrimination (Roberson &
Jonathan C. Ziegert and Paul J. Hanges, Department of Psychology,
University of Maryland.
We thank Brent Smith and Susan Taylor for their comments on drafts of
this article. An earlier version of this article was presented at the 17th
Annual Conference of the Society for Industrial and Organizational Psy-
chology, April 2002, Toronto, Ontario, Canada.
Correspondence concerning this article should be addressed to Jonathan
C. Ziegert, Department of Psychology, University of Maryland, College
Park, MD 20742. E-mail: jziegert@psyc.umd.edu
Journal of Applied Psychology Copyright 2005 by the American Psychological Association
2005, Vol. 90, No. 3, 553–562 0021-9010/05/$12.00 DOI: 10.1037/0021-9010.90.3.553
553
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Block, 2001). In the present study, we adopted the disparate
treatment definition of employment discrimination; disparate treat-
ment occurs when different standards are applied to different
groups (Gatewood & Field, 2000). Several meta-analyses have
been conducted to identify the factors that impact the magnitude
and direction of these differential standards. For example, meta-
analyses by Ford, Kraiger, and Schechtman (1986) and Roth,
Huffcutt, and Bobko (2003) have found that Blacks received lower
scores and evaluations on both objective and subjective measures.
In addition, Kraiger and Ford (1985) found that both Black and
White raters gave higher ratings to members of their own race.
These results provide evidence that individuals apply differential
standards when evaluating applicants. Indeed, Roth et al. (2003)
have called for future research to identify the biases that contribute
to these racial differences in performance.
Of course, mean differences between Black and White appli-
cants on some measurement instrument do not necessarily imply
bias. It is possible that the mean differences reflect true differences
between groups (Oppler, Campbell, Pulakos, & Borman, 1992). To
address this possibility, Oppler et al. (1992) controlled for a
number of job relevant variables and found that although the
differences between Blacks and Whites diminished, they did not
disappear. Recognizing the fact that it is unlikely that the authors
included all of the relevant variables that might explain these
differences, the Oppler et al. study does provide indirect evidence
that racial attitudes might be a contributing factor to the average
rating differences for Blacks and Whites.
A series of studies by Brief et al. (2000) provides more direct
evidence for the role of racial attitudes. In the second study
reported in the Brief et al. (2000) article, participants completed
the Modern Racism Scale (MRS; McConahay et al., 1981) as well
as an in-basket exercise. Among the many in-basket items was a
selection decision task that asked the participants to rate potential
job applicants. In the in-basket background material, participants
in one condition received a memo from the simulated organiza-
tion’s president that contained a statement expressing his desire for
the candidate hired to be White. Another condition received a
memo from the president, but it did not contain the statement
concerning racial preferences. Although Brief et al. did not discuss
this “social-norm” manipulation as creating an organizational cli-
mate for racial bias or equality, their manipulation is consistent
with the conceptualization of climate proposed by Schneider
(1972). Schneider indicated that organizational climate is a func-
tion of what is rewarded, supported, and expected in the organi-
zation and sends strong signals to employees and others about
what behavior is socially acceptable.
As expected, Brief et al. (2000) found that the average rating of
Black applicants was lower in the climate for racial bias condition
(when the president was perceived as a legitimate authority and
indicated his White racial preference). More important, they found
that scores on the MRS moderated this relationship in that modern
racists gave Black applicants lower ratings in the climate for racial
bias condition. Thus, it appears that modern racists act on their
beliefs but only when the social norms appear to legitimize dis-
crimination. However, before accepting this interpretation, it is
important to note that Brief et al. (2000) operationalized biased
ratings as a significant difference in the average evaluation of
Black applicants across the two climate conditions. In other words,
they did not assess what happened to the ratings of White appli-
cants across the two climate conditions. It is possible that the
biased climate condition universally decreased the ratings of all
applicants, regardless of their race. Thus, in the present study, we
attempt to replicate this effect for organizational climate through
the use of a slightly more sensitive measure of discrimination
through comparisons of the differential rating of Black and White
applicants across the two climate conditions:
Hypothesis 1 (H1): Organizational climate will influence ap-
plicant ratings. In particular, Black applicants will have lower
ratings compared with White applicants in the “climate for
racial bias” condition. The magnitude of the rating differences
between Black and White applicants will be smaller in the
“climate for equality” condition.
Old-Fashioned and Modern Racism Explicit Measures
Stereotypes and prejudice can manifest themselves in different
ways. One of the most obvious and salient forms is old-fashioned
racism, which is characteristic of blatant attitudes of the inferiority
of Blacks and open bigotry. In contrast to these openly espoused
racial prejudices, modern racism has evolved as a newer and
subtler form of racism (McConahay, 1986). Modern racism is
more indirect and rationalized where negative attitudes toward
Blacks are masked with nonracial reasons to preserve a nonpreju-
dicial self-image (Brief, 1998). The central tenets and beliefs of
modern racists include the thinking that discrimination is a thing of
the past, Blacks are using unfair tactics to push themselves into
places where they are not wanted, and gains by Blacks are not
deserved (McConahay, 1986). Modern racists see their beliefs as
constituting empirical facts (Brief, 1998; McConahay, 1986). They
do not believe that they are racist because they conceptualize a
racist as someone who espouses the old-fashioned or hostile rac-
ism beliefs. As illustrated by Brief et al. (2000), modern racists
will act on their beliefs when there is some social norm (e.g.,
climate for racial bias) justifying their discriminatory behaviors. In
the present study, we aim to constructively replicate this Brief et
al. finding by using our comparative definition of discrimination as
well as measuring old-fashioned and modern racism:
Hypothesis 2 (H2): Organizational climate will interact with
participants’ racist attitudes (old-fashioned and modern) to
affect the average ratings of Black compared with White
applicants. Specifically, the more racist participants (as mea-
sured by the explicit old-fashioned and modern racism mea-
sures) will produce more discriminatory ratings in the “cli-
mate for racial bias” condition. There should be no
relationship between the explicit racism measures and aver-
age applicant ratings in the “climate for equality” condition.
As stated in H2, we expect that the direction of the results will be
similar for both the old-fashioned and modern racism scales.
However, we expect that the effect will be weaker for the old-
fashioned measure because of the blatant, extreme nature of the
items on this scale.
Implicit Racist Attitudes
As discussed above, the most common way to measure racist
attitudes is through explicit self-report measures. Unfortunately,
554
RESEARCH REPORTS
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
such measures are influenced by self-presentation biases in which
respondents’ scores reflect not only their attitudes but also the
deliberate and conscious manipulation of responses to regulate
their impression to others (Dunton & Fazio, 1997; Plant & Devine,
1998). One way to minimize the influence of self-presentation bias
in the measurement of racist attitudes is through the inclusion of
implicit attitude measurement techniques. These implicit measure-
ment techniques seek to indirectly assess a construct without
having to directly ask participants for a verbal report (Fazio &
Olson, 2003). Thus, implicit measures assess attitudes in a more
subtle fashion than do more traditional self-report measures. In-
deed, the discrepancy between prejudicial attitudes assessed with
implicit and explicit measures has led to questions of the trustwor-
thiness of easily monitored explicit responses such as verbal re-
ports and self-report ratings (Fazio et al., 1995).
As the use of implicit measures to assess attitudes has grown in
recent years, so have questions regarding their meaning and va-
lidity. For example, Crandall and Eshleman (2003) have noted that
although implicit attitude measurement holds promise, they also
questioned the degree to which these measures reflect genuine
prejudice, and they suggested that “genuine prejudice and implicit
attitudes are related, but they are not the same concept” (p. 435).
Fazio and Olson (2003) recently reviewed several types of implicit
measurement techniques and concluded that “implicit measures
have the potential to serve as useful methodological tools for
testing hypotheses” (p. 320). Given this initial evidence for the
validity and potential utility of implicit measurement methods,
there are several reasons, besides the freedom from self-
presentation biases, why implicit attitude measures may be more
useful than explicit attitude measures. For example, implicit atti-
tude measurement techniques are believed to reflect the more
ingrained beliefs of the respondents. These ingrained beliefs may
be activated automatically outside of the person’s consciousness
(Greenwald & Banaji, 1995), and thus, they might influence be-
havior more than explicit attitudes that are produced as a result of
conscious, deliberate processing (Bargh & Chartrand, 1999). An-
other possible advantage is that individuals may not be able to
adequately access their introspective processes (Nisbett & Wilson,
1977). Thus, attitudes measured implicitly should be more predic-
tive of behavior than attitudes measured explicitly. Therefore, a
logical extension of H2 is to examine it with implicit attitudes:
Hypothesis 3 (H3): The relationship between individuals’
implicit racist attitudes and the ratings of Black applicants
(compared with the ratings of White applicants) will be
moderated by the climate condition. In particular, participants
who hold racist views measured implicitly will act in a
discriminatory fashion in the climate for racial bias condition,
but these individuals should not exhibit any discriminatory
behavior in the climate for equality condition.
H3 is identical to H2 except that it substitutes implicit for explicit
attitude measurement. On the basis of the assumption that implicit
attitudes are more advantageous than are explicit attitudes, this
interaction should be stronger for the implicit rather than explicit
measures.
In the current study, we measured implicit attitudes with the
Implicit Association Test (IAT) developed by Greenwald,
McGhee, and Schwartz (1998). The IAT is a flexible tool that can
be readily adapted to measure a variety of concepts. It is admin-
istered via a computer and it assesses implicit attitudes by mea-
suring the response latency and errors associated with individuals
sorting words from four concepts into only two categories. In the
current study, we used the same concepts, categories, and words as
in the Greenwald et al. (1998) race IAT measure.
There is some debate regarding the meaningfulness of this
measure (e.g., Devine, 2001). For example, it has been suggested
that shifts in response criteria can impact IAT scores (Brendl,
Markman, & Messner, 2001) or that the IAT reflects the environ-
ment that a person is a part of as opposed to that person’s endorsed
attitudes (Karpinski & Hilton, 2001). The IAT seems to be sus-
ceptible to in-group/out-group distinctions, suggesting the auto-
matic nature of intergroup bias (Ashburn-Nardo, Voils, and Mon-
teith, 2001). Research also illustrates that manipulations in the
situation or context can impact the magnitude of results (e.g.,
Dasgupta & Greenwald, 2001).
Although these studies document factors that affect the IAT,
there is also a great deal of validity evidence for it. Greenwald and
Nosek (2001) reviewed over 30 studies that discussed the psycho-
metric and validity evidence for the IAT. Research has shown that
the IAT can detect stable attitude differences (Greenwald et al.,
1998), and the implicit attitude score is not influenced by famil-
iarity of the words used in the task (Dasgupta, McGhee, Green-
wald, & Banaji, 2000). Further, the IAT has shown convergent
validity with other latency measures and priming tasks (Rudman &
Kilianski, 2000) as well as with a physiological measure (Phelps et
al., 2000). Finally, there is also support for predictive validity of
the IAT in predicting more subtle behavior such as nonverbal
behavior (McConnell & Leibold, 2001). Overall, there has been
much research that has provided evidence for the psychometric
properties and the internal, convergent, discriminant, and predic-
tive validity of the IAT (Greenwald & Nosek, 2001). The present
study adds to this validity evidence by assessing whether the IAT
predicts more macrolevel behavior such as discriminatory hiring
decisions.
Motivation to Control Prejudice
One possible explanation for the differences between implicit
and explicit attitudes is that people typically do not like to be
thought of as prejudiced. In other words, some people are moti-
vated to control their prejudice. Dunton and Fazio (1997) have
shown that implicitly prejudiced individuals with high motivation
to control their prejudice will generally respond in a nonprejudiced
way on self-report (i.e., explicit) measures. Thus, it is reasonable
to expect that there will be less association between the explicit
and implicit racism measures when people are motivated to control
their prejudices.
Hypothesis 4 (H4): The relationship between implicit and
explicit racial attitudes will be moderated by motivation to
control prejudice. Specifically, the stronger the implicit prej-
udicial attitudes and the stronger the motivation to control
prejudice, the more negative the relationship will be between
implicit and explicit attitudes.
This relationship is expected to be stronger for old-fashioned
rather than modern racist attitudes as modern racism is purported
to be subtler and less susceptible to presentational biases.
555
RESEARCH REPORTS
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Method
Participants
Participants were 103 undergraduates from a large mid-Atlantic public
university who received course credit. Ages ranged from 18 to 30 with a
mean of 18.8 (SD 1.5) years. The majority of the participants were
women (61.2%; men 38.8%), and because the purpose of the experiment
was to assess bias against Blacks, all participants were non-Black with
Whites as the majority (White 81%; Asian 7%; Latino/Hispanic
6%; Arab 2%; and Other 4%). There were no significant differences
among these groups for any of the study variables.
Measures
Explicit racial attitudes. Racism was measured explicitly with two
self-report scales: the Attitudes Toward Blacks Scale (ATB; Brigham,
1993) and the MRS. The ATB is a 20-item measure of old-fashioned racist
attitudes on a 7-point Likert-type scale. For example, participants are asked
if they would dislike living near Black people and if they feel that Black
and White people are equal. The reliability of participants’ scores on this
scale was acceptable with a coefficient alpha of .89. In contrast to the ATB,
the MRS assesses modern racism. The MRS contains seven items mea-
sured on a 7-point Likert-type scale. An example item is “Discrimination
against Blacks is no longer a problem in the United States.” The modern
racism items were embedded with 14 other items assessing attitudes toward
other issues (e.g., homosexuality and abortion) to limit potential reactivity
effects. The reliability of the scores of this scale was acceptable with a
coefficient alpha of .81.
Implicit racial attitudes. Implicit racism was assessed with the IAT.
The IAT measures racist attitudes by recording the speed and accuracy
with which participants can categorize words. Participants first sorted
words into a descriptive category relevant to race (i.e., names that Green-
wald et al. [1998] had determined to be perceived as prototypical of Whites
or Blacks) and then sorted the next group of words into an evaluative
category (i.e., pleasant vs. unpleasant words). In particular, names ap-
peared on the computer screen one at a time, and participants had to
categorize these names as being either “Black” or “White” using the left or
right key on the computer keyboard. In the next trial, words appeared on the
computer screen one at a time and participants had to categorize them as being
either “pleasant” or “unpleasant.” These categorization tasks were then com-
bined and participants had to categorize 40 words as either belonging to the
“White or pleasant” category or the “Black or unpleasant” category. The speed
and accuracy of the categorization process were recorded. After this combined
task trial, the race categories were switched to allow for practice and were then
combined with the evaluative category. That is, participants determined
whether a word belonged to the “White or unpleasant” category versus the
“Black or pleasant” category for another 40 words.
1
An example of this
sequence of trials for the IAT can be seen in Figure 1.
We focused on both the speed (response latency) as well as accuracy
(percentage of errors) in the word classifications as our assessment of
implicit attitudes. We examined the degree to which participants were
slower in their responding and made more errors (initially placed the word
in the wrong category) when the “White and unpleasant” categories were
paired than when “White and pleasant” were paired.
2
In particular, laten
-
cies and percentage of errors for the “White and pleasant” trial were
subtracted from the “White and unpleasant” trial, which results in positive
latencies and percentage of errors being indicative of implicit racist atti-
tudes (individuals can associate “White and pleasant” faster and make less
errors than they can with “Black and pleasant”). We followed the proce-
dure of Greenwald et al. (1998) and recoded responses below 300 ms and
above 3,000 ms, and we eliminated cases with an average error rate of over
20% (indicative of random responding), which resulted in only a small
number of cases being deleted (less than 5% of the sample; after deletion
of these cases, the final sample was 103 participants).
Average response latencies ranged from 368.03 ms to 739.15 ms with
a mean of 233.52 ms (SD 168.36). This positive value indicates that the
average implicit racial attitude for our sample was somewhat negatively
biased against Blacks (e.g., faster classification for the “White and pleas-
ant” than for the “White and unpleasant”). The error rate percentage ranged
from 10% to 15% with an average of 1.5% (SD 4.63), which also
indicates that the average implicit attitude was somewhat negatively biased
against Blacks.
We subjected these two measures to an exploratory principal compo-
nents factor analysis, and by examining the eigenvalues we found evidence
that only one factor emerged. Both scales had factor loadings of .79 with
this single factor. Using these factor loadings and the reliability formula
provided by Bollen (1989), we estimated the reliability of each scale to be
.63. As these two measures are on different metrics, we created a composite
implicit racist attitude score by using z scores to standardize the response
latency and the error rate percentage and then by adding these two z scores
together. Using the linear composite reliability formula provided by Nun-
nally and Bernstein (1994), we estimated the reliability of the composite
implicit measure to be .70.
Motivation to control prejudice. Motivation to control prejudice was
assessed through the use of Plant and Devine’s (1998) External subscale on
the Motivation to Respond Without Prejudice measure. This External
subscale measures motivation to hide racial prejudices in order to conform
to societal norms and appear nonprejudiced to others. The five items for
this measure were responded to on a 9-point Likert-type scale. An example
item is “I try to hide any negative thoughts about Black people in order to
avoid negative reactions from others.” The reliability of participants’
scores on this scale was acceptable with a coefficient alpha of .77. Validity
evidence and freedom from social desirability concerns for this scale have
been demonstrated on the basis of small correlations with several social
desirability scales (Plant & Devine, 1998).
Experimental task and racial discrimination measure. Participants
completed the in-basket exercise used by Brief et al. (2000). It contained
many tasks typically encountered by managers (e.g., determining compen-
sation for a newly hired employee). Embedded in this in-basket was a
“hiring recommendation” task. Participants were provided with the dos-
siers of eight job applicants and were instructed to evaluate them. The
dossiers provided information about each applicant’s education, prior work
experience, race, gender, and hobbies. These dossiers were written such
that six of the eight applicants had outstanding qualifications. Participants
rated each applicant to the degree to which they were an exceptional
referral on a 5-point Likert-type scale ranging from 1 (should not have been
referred)to5(excellent referral). Prior work has shown that there are no
differences among these six candidates when race information is removed
(Brief, Buttram, Elliott, Reizenstein, & McCline, 1995). The race of these
applicants was randomly assigned with half (three) of the qualified appli-
cants as Black and the other half (three) as White. In addition to race, the
sex of the applicants was also randomly assigned so that one each of the
qualified Black and White candidates was a woman (i.e., there were two
Black men, two White men, one Black woman, and one White woman).
1
This order was counterbalanced such that half of the participants first
completed the pairing of “White or pleasant” versus “Black or unpleasant,”
whereas the other half of the participants first completed “White or un-
pleasant” versus “Black or pleasant.” Further, the order of the specific
words that appeared was randomized. Finally, consistent with Greenwald
et al. (1998), there was one practice trial that took place before each of the
combined categorization tasks in order to familiarize participants with the
task. Data were not recorded for these practice trials.
2
Alternatively, because of the nature of the measure, this could have
been equally phrased regarding the degree to which participants were
slower in their responding and made more errors when “Black and pleas-
ant” categories are paired than when “Black and unpleasant” are paired.
556
RESEARCH REPORTS
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
In their second study, Brief and colleagues (2000) used the average
rating of the three qualified Black applicants as their measure of racial
discrimination. Our measure was obtained by conducting within-individual
regression analyses through the use of a dummy-coded race variable (i.e.,
0 White applicant; 1 Black applicant) to predict a participant’s ratings
of the six applicants. More specifically, a hierarchical linear model (HLM)
was conducted with the dummy-coded race variable used as a Level 1 (i.e.,
within-individual) predictor. The slopes for this dummy-coded variable
were used as our dependent measure of racial discrimination. A negative
slope was indicative of bias against the Black applicants (i.e., the mean for
the Black applicants was lower than the mean for the White applicants), a
zero slope was indicative of no racial bias (i.e., equal mean ratings of Black
and White applicants), and a positive slope was indicative of a pro-Black
applicant bias (i.e., the mean for the Black applicants was higher than the
mean for the White applicants). The use of the within-individual slope as
our assessment of racial discrimination produced a sensitive measure
because it accounts for any disparate ratings between Black and White
applicants.
Procedure
At the beginning of the semester, introductory psychology students
participated in a mass testing session in which they were asked to complete
a variety of measures. Included were the two explicit racist attitude
measures (ATB and MRS) as well as the Motivation to Respond Without
Prejudice Scale. These measures were spaced equally among a number of
Figure 1. Sample illustration of the Implicit Association Test (IAT). Participants completed a series of five
trials in which the target concepts and attributes were introduced in the first two trials. The targets and attributes
were combined in the third trial and reversed in the fifth trial after initially reversing the target categories in the
fourth trial. The correct categorization of the sample stimuli that appears one at a time is illustrated with a check
mark. Implicit racist attitudes exist if a participant takes longer and makes more errors when “Black or pleasant”
(also “White or unpleasant”) are paired than when “Black or unpleasant” (also “White or pleasant”) are paired.
557
RESEARCH REPORTS
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
other scales in the mass testing session to minimize any response con-
founds from potentially completing the measures sequentially. Approxi-
mately 1 month later, participants were recruited from the mass testing
sample for the laboratory portion of the study.
Upon arrival, participants were asked to play the role of a manager and
complete the in-basket exercise. Participants were randomly assigned to
either the climate for equality or the climate for racial bias condition. This
manipulation took place through a memo from the president of the com-
pany. In both conditions, the president instructed individuals to take into
account applicants’ education and experience in making their evaluations
of each candidate. However, for participants in the climate for racial bias
condition, the president’s memo also stated the following:
Given that the vast majority of our workforce is White, it is essential
we put a White person in the VP position. I do not want to jeopardize
the fine relationship we have with our people in the units. Betty (the
outgoing vice president) worked long and hard to get those folks to
trust us, and I do not want her replacement to have to overcome any
personal barriers.
Participants in the climate for equality condition did not receive this
statement. After completion of the in-basket and the embedded hiring
recommendation task, participants completed a manipulation-check item,
the implicit measure of racism, and a demographic questionnaire.
Manipulation check. In order to determine whether participants were
cognizant of the instructions in each condition, participants were asked to
recall the hiring preferences of the president. The item asked whether the
president preferred to hire applicants that were White, Black, Latino, or of
no stated preference.
Statistical analyses. We tested the first three hypotheses with a random
slope HLM analysis and the fourth with moderated regression. The hy-
potheses using HLM were tested through the use of between-person (i.e.,
Level 2) variables to predict the magnitude of the within-person slope
measure of racial discrimination.
Results
Manipulation Check and Preliminary Evidence of Racial
Discrimination
Analysis of the manipulation check indicated the hiring prefer-
ences of the president were correctly recalled as 95.15% of the
participants identified the proper preference for their respective
condition; this indicates that the climate for racial bias versus
climate for equality manipulation worked. We first performed an
HLM analysis to assess the extent to which ratings were biased.
With the within-person Black–White slope (i.e., the Level 1 pre-
dictor was the dummy-coded race variable) as our measure of
racial bias and no Level 2 predictors, results indicated an overall
bias against Blacks; that is, the Black applicants were rated lower
than were the White applicants (i.e., R
2
within
.26, b
y.x
⫽⫺.58,
t[96] ⫽⫺7.25, p .05).
3
We also examined the variance of this
slope and found that it was significantly different from zero (i.e.,
2
Slope
0.27,
2
(96, N 99) 172.92, p .05), which
indicates that some of the participants exhibited more bias than did
others.
4
As Black applicants were rated lower and there were
meaningful individual differences in this degree of bias, we pro-
ceeded to test our hypotheses in order to explain these slope
differences.
Test of Hypotheses
Table 1 presents the means, standard deviations, and intercor-
relations among the study variables. H1 predicted that participants
in the climate for racial bias condition would exhibit greater
discrimination than would participants in the climate for equality
condition. To test H1, we ran a randomized slope HLM model in
which we predicted the Level 1 Black–White slope through the use
of a Level 2 dummy-coded corporate climate predictor (i.e., 0
climate for equality;1 climate for racial bias). The Level 2
3
As indicated earlier, the in-basket task contained six exceptional can
-
didates: two Black men, one Black woman, two White men, one White
woman. Given that the applicants differed on race and gender, it is possible
that the gender of the applicant interacted with applicant race to influence
participant’s ratings. To test whether race and gender interacted, we per-
formed an HLM analysis in which the race, gender, and the interaction of
these two variables were used to predict participant ratings. The analyses
revealed that the Race Gender interaction was not significant (R
2
within
.00, b
y.x
⫽⫺.11, t[96] ⫽⫺1.00, ns).
4
The HLM analyses indicated that there was a significant applicant
gender effect in that male applicants were rated significantly higher than
were female applicants (R
2
within
.05, b
y.x
⫽⫺.28, t[96] ⫽⫺4.87, p
.05). Unlike the race of the applicant, however, this gender effect did not
significantly differ across participants (
2
Slope
0.02,
2
[96, N 99]
86.72, ns).
Table 1
Means, Standard Deviations, and Intercorrelations Among Predictor Variables
Variable MSD 12 3 4567
1. Attitude Toward Blacks (ATB) Scale 2.64 0.88
2. Modern Racism Scale (MRS) 2.79 1.05 .73**
3. Implicit z-score
a
composite
0.00 1.59 .12 .12
4. Motivation to control prejudice 4.98 1.62 .06 .17 .07
5. Mean rating of White applicants
c
4.30 0.53 .15 .22* .02 .03
6. Mean rating of Black applicants
c
3.73 0.74 .19 .18 .29** .02 .29**
7. Climate condition
b
0.48 0.50 .10 .12 .12 .02 .23* .25*
Note. As a result of missing data, n ranges from 99–101. Higher scores equal greater amounts of the construct.
a
The implicit measure of racism is a standardized z-score composite, thus the mean is 0. This sample indicated implicit racist attitudes on the basis of the
unstandardized scores.
b
For the climate condition, 0 climate for equality and 1 climate for racial bias.
c
Although our measure of discrimination
was the slope obtained by conducting within-individual regression analyses using a dummy-coded race variable of the ratings of the six applicants, we have
provided the mean ratings of the White and Black applicants for illustrative purposes.
* p .05. ** p .01.
558
RESEARCH REPORTS
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
climate variable significantly predicted the magnitude of the
Black–White slope (R
2
between
.33,
⫽⫺.50, t(95) ⫽⫺3.57,
p .05). Consistent with Brief et al. (2000), participants in the
climate for racial bias condition exhibited a stronger degree of bias
than did participants in the climate for equality condition. An
examination of the within-person Black–White slope indicates that
there was still a significant amount of between-person slope vari-
ance (
2
⫽⫺.18,
2
[95] 146.52, p .05), even after account
-
ing for the corporate climate manipulation.
H2 predicted that the relationship between participants’ explicit
racist attitudes and discriminatory ratings would be moderated by
the corporate climate manipulation such that individuals would
exhibit more discrimination against Black applicants as their de-
gree of explicit racial attitudes increased, but only in the climate
for racial bias condition. H2 was tested for the ATB and MRS
separately. With respect to the ATB, a random slope HLM model
was conducted in which the Level 1 Black–White slope was
predicted by using the Level 2 dummy-coded climate predictor, the
ATB, and the interaction between these two variables. Contrary to
H2, the interaction term was not significant (R
2
between
.01,
.20, t[93] ⫽⫺1.18, ns). Similarly, the random slope HLM with
the MRS failed to find a significant interaction between the cor-
porate climate manipulation and the MRS (R
2
between
.00,
.10, t[93] ⫽⫺0.91, ns). As H2 was not supported, we failed to
replicate the interaction reported in Brief et al. (2000).
H3 was similar to H2 except that it explored the utility of the
implicit attitude measure rather than the explicit measures. As
predicted, HLM analysis indicated that the IAT z-score composite
significantly interacted with the climate condition (R
2
between
.05,
⫽⫺.17, t[93] ⫽⫺2.03, p .05),
5
and this interaction is
shown in Figure 2. The amount of discrimination against Blacks
was greater for participants with more implicitly racist attitudes in
the climate for racial bias condition. The relationship between
discrimination and implicit racist attitudes was almost nonexistent,
with only a slight upward trend in the climate for equality condi-
tion. In summary, there was support for H3; the climate for racial
bias manipulation sent a signal about social norms for discrimina-
tion that resulted in implicit racial attitudes being related to dis-
parate ratings of Black and White applicants.
H4 predicted that the relationship between explicitly and im-
plicitly measured racial attitudes would be moderated by motiva-
tion to control prejudice. Hierarchical multiple regressions illus-
trated that the predicted interaction was significant for the ATB
scale (R
2
.05, F
inc
(1, 95) 4.76, p .05), although it did not
reach significance for the MRS (R
2
.02, F
inc
(1, 95) 2.09, ns;
see Table 2). This significant interaction for the ATB is depicted
in Figure 3. This interaction is consistent with the hypothesized
direction; that is, when individuals have low motivation to control
their prejudice, there is a positive relationship between the implicit
5
The final Level 2 equation consisting of the condition variable, the
implicit racism measure, and the interaction between these measures ac-
counted for 46% of the between-person variance.
Figure 2. Relationship of Implicit Association Test (IAT) z-score composite and slope for race moderated by
corporate climate condition.
559
RESEARCH REPORTS
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
and explicit racism measures. However, when individuals are
highly motivated to control their prejudice, there is a slight nega-
tive relationship between their implicit and explicit racism mea-
sures. When highly motivated to hide their prejudice, individuals
holding implicit racist attitudes tend to distort their explicit racism
responses more in order to report less explicit racial stereotypes.
Although we originally hypothesized that both the ATB as well as
the MRS would show an effect, the fact that only the ATB
interaction was significant is not surprising as this scale taps a
more old-fashioned racism, which is easier to monitor and control
than is modern racism.
Discussion
The purpose of this study was to constructively replicate and
extend the research conducted by Brief et al. (2000) that con-
cerned employment discrimination by using a scenario-based
laboratory study with undergraduate students. Our study built
upon this prior work in several ways in order to seek a more
detailed understanding of the antecedents of employment dis-
crimination. First, we included a measure of old-fashioned
explicit racist attitudes in addition to modern racism. The
old-fashioned explicit racism measure was included to ascertain
whether modern racism scales were really needed to identify the
individuals in our study that would act in a prejudicial manner.
Second, we included an implicit racial attitude measure in
addition to these two explicit attitude measures. Thus, there was
a graded distinction between explicit and implicit racist atti-
tudes from old-fashioned (ATB), to modern (MRS), to implicit
(IAT). Third, we included a measure of motivation to control
prejudice to directly test whether it is a self-presentation bias
that accounts for the different results obtained by explicit and
implicit measures. Finally, we used HLM to produce a more
sensitive measure of racial discrimination as this measure di-
rectly compares differences in the ratings of Blacks and Whites.
Taken together, these additions allowed us to constructively
replicate Brief et al.’s study and to test new hypotheses to
provide a greater understanding of employment discrimination.
Figure 3. Relationship of implicit (Implicit Association Test [IAT] z-score composite) and explicit (Attitude
Toward Blacks Scale) attitudes moderated by motivation to control prejudice.
Table 2
Moderated Regression Analyses of Explicit Racist Attitudes and
Motivation to Control Prejudice on Implicit Racist Attitudes
Variable
SE R
2
F
Modern racism
Modern racism (A) 0.131 .166 .02 1.02
Motivation to control
prejudice (B) 0.088 .101
A B 0.561 .102 .02 2.09
Attitude toward Blacks
Attitude toward
Blacks (A) 0.112 .196 .02 0.81
Motivation to control
prejudice (B) 0.056 .100
A B 1.103* .128 .05* 4.76*
* p .05.
560
RESEARCH REPORTS
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
The results help to illustrate the differences between implicit
and explicit attitudes in understanding bias in organizations. We
were unable to replicate Brief et al.’s (2000) interaction between
explicit racist attitudes and the climate for racial bias manipulation.
However, there are several potential explanations for why we were
unable to find similar results. For example, we used a different
measure of discrimination than did Brief et al. as we compared
ratings of both White and Black applicants instead of just assessing
ratings of Black applicants. This comparison provides a different
type of assessment of discrimination that focuses on disparate
treatment, which may have led to the discrepant results. Another
potential explanation may be that the sample for the current study
reported less racist attitudes than did the sample in the Brief et al.
study (as measured by the MRS). It could be the case that our
sample was less racist than was the earlier sample. Alternatively,
the topic of race has become even more salient in recent years;
thus, it may be that the current sample exhibited more self-
presentational forces in responding to the explicit measures of
racism, which masked any effects linking self-reported racist
views on questionnaires to discriminatory behavior. This potential
for self-presentational forces contributing to the discrepant find-
ings is strengthened by the fact that motivation to control prejudice
moderated the relationship between one of the explicit measures
(ATB) and the implicit measure of racism.
Although we were unable to replicate this finding with explicit
attitudes, we did find a significant interaction through the use of
implicit attitudes. More specifically, implicit racism interacted
with a climate for racial bias to predict discrimination; when
individuals were given a business justification for racial discrim-
ination their implicit racist attitudes were positively related to their
discriminatory behavior. There are several implications of these
laboratory-based findings with an undergraduate student sample.
First, although we could not directly replicate Brief et al.’s (2000)
findings with the explicit racism measures, this interaction con-
ceptually replicates the finding that subtle biases will be acted
upon in the right social environment. Second, our results demon-
strate that implicit attitudes can be used to predict meaningful
macrolevel behavior. Though there have been a few studies relat-
ing the IAT to behavior (e.g., McConnell & Leibold, 2001), these
studies tend to relate the IAT to microlevel behaviors (e.g., less
speaking and smiling with a Black experimenter) or to survey
responses. This is one of the first studies to demonstrate that the
IAT can predict racially biased discriminatory actions. Thus, this
study highlights the usefulness of assessing attitudes with an
implicit technique and the potential importance of using it to
predict discrimination.
Finally, the finding that the relationship between implicit and
explicit attitudes is moderated by the motivation to control preju-
dice is of interest because it verifies the hypothesized mechanism
by which implicit and explicit attitudes are related to one another.
When researchers use an explicit measurement technique, individ-
uals who are motivated to control their responses may indeed do
so. Conversely, individuals do not seem to monitor and change
their responses when attitudes are assessed implicitly. This finding
lends support to the assertion that explicit attitude measurement
may be prone to self-presentational forces, whereas implicit tech-
niques are not.
It should be noted that these hypotheses do not depend on
whether individuals can accurately assess their implicit attitudes.
There is some question as to the degree to which implicit attitudes
are available to introspection (Fazio & Olson, 2003). Regardless,
these findings illustrate the effects of implicit attitudes and their
difference from explicit attitudes. Results show that whether or not
individuals can accurately assess their own implicit attitudes, these
implicit attitudes are indeed predictive of behavior and are differ-
ent than explicit attitudes.
As with any study, the current research has several limitations.
As a laboratory study conducted with undergraduates, several
generalizability issues arise in the interpretation of results. For
example, the use of “paper” candidates may be problematic as it
raises questions of external validity. Further, as the participants
were students, they may have had very little experience in the type
of role (manager) that they played during the in-basket exercise.
Though these limitations are inherent to this laboratory study, we
felt that it was necessary to initially study this phenomenon in the
lab to obtain a high degree of control so that the candidates being
rated were the same except for race and that an accurate measure-
ment of implicit attitudes was obtained. Even with these potential
limitations in generalizability, we believe the current study makes
an important contribution to the literature.
The present findings have both theoretical and practical impli-
cations. Implicit attitudes are not only different than are self-
reported explicit attitudes, but the present research illustrates that
implicit attitudes may be more predictive of behavior in certain
situations. Therefore, it is important to recognize the difference
between implicit and explicit attitudes and the resulting impact on
behavior. In terms of practical contributions, our results suggest
that when trying to understand the reactions and behaviors of
individuals who are in a position to discriminate against certain
groups of individuals, one cannot rely on explicitly espoused
attitudes alone, but instead one may need to understand the char-
acteristics of the situation, the motivation of the individuals to
conceal their prejudice, and certain implicit attitudes.
The current study extends the work by Brief et al. (2000) by
illustrating that implicit attitudes are important components in
understanding employee discrimination. This study shows that
motivation to control prejudice can be used to potentially explain
the distinction between implicit and explicit attitudes. It also
provides several reasons why implicit attitudes may be ultimately
more useful than explicit attitudes. The message that emerges from
these findings suggests that as employment discrimination has
changed in recent years, so should the conceptualization of the
attitudes used to predict it.
References
Ashburn-Nardo, L., Voils, C. I., & Monteith, M. J. (2001). Implicit
associations as the seeds of intergroup bias: How easily do they take
root? Journal of Personality and Social Psychology, 81, 789 –799.
Bargh, J. A., & Chartrand, T. (1999). The unbearable automaticity of
being. American Psychologist, 54, 462–479.
Bollen, K. A. (1989). Structural equations with latent variables. New
York: Wiley.
Brendl, C. M., Markman, A. B., & Messner, C. (2001). How do indirect
measures of evaluation work? Evaluating the inference of prejudice in
the Implicit Association Test. Journal of Personality and Social Psy-
chology, 81, 760 –773.
Brief, A. P. (1998). Attitudes in and around organizations. Thousand Oaks,
CA: Sage.
561
RESEARCH REPORTS
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Brief, A. P., Buttram, R. T., Elliott, J. D., Reizenstein, R. M., & McCline,
R. L. (1995). Releasing the beast: A study of compliance with orders to
use race as a selection criterion. Journal of Social Issues, 51, 171–193.
Brief, A. P., Dietz, J., Cohen, R. R., Pugh, S. D., & Vaslow, J. B. (2000).
Just doing business: Modern racism and obedience to authority as
explanations for employment discrimination. Organizational Behavior
and Human Decision Processes, 81, 72–97.
Brigham, J. C. (1993). College students’ racial attitudes. Journal of Applied
and Social Psychology, 23, 1933–1967.
Bryk, A. P., & Raudenbush, S. W. (1992). Hierarchical linear models.
Newbury Park, CA: Sage.
Crandall, C. S., & Eshleman, A. (2003). A justification-suppression model
of the expression and experience of prejudice. Psychological Bulletin,
129, 414 446.
Dasgupta, N., & Greenwald, A. G. (2001). On the malleability of automatic
attitudes: Combating automatic prejudice with images of admired and
disliked individuals. Journal of Personality and Social Psychology, 81,
800 814.
Dasgupta, N., McGhee, D. E., Greenwald, A. G., & Banaji, M. R. (2000).
Automatic preference for White Americans: Eliminating the familiarity
explanation. Journal of Experimental Social Psychology, 36, 316 –328.
Devine, P. G. (2001). Implicit prejudice and stereotyping: How automatic
are they? Introduction to the special section. Journal of Personality and
Social Psychology, 81, 757–759.
Dunton, B. C., & Fazio, R. H. (1997). An individual difference measure of
motivation to control prejudiced reactions. Personality and Social Psy-
chology Bulletin, 23, 316 –326.
Fazio, R. H., Jackson, J. R., Dunton, B. C., & Williams, C. J. (1995).
Variability in automatic activation as an unobtrusive measure of racial
attitudes: A bona fide pipeline? Journal of Personality and Social
Psychology, 69, 1013–1027.
Fazio, R. H., & Olson, M. A. (2003). Implicit measures in social cognition
research: Their meaning and use. Annual Review of Psychology, 54,
297–327.
Ford, J. K., Kraiger, K., & Schechtman, S. L. (1986). Study of race effects
in objective indices and subjective evaluations of performance: A meta-
analysis of performance criteria. Psychological Bulletin, 99, 330 –337.
Gatewood, R. D., & Field, H. S. (2001). Human resource selection (5th
ed.). Stamford, CT: Thomson Learning.
Greenwald, A. G., & Banaji, M. R. (1995). Implicit social cognition:
Attitudes, self-esteem, and stereotypes. Psychological Review, 102,
4 –27.
Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring
individual differences in implicit cognition: The Implicit Association
Test. Journal of Personality and Social Psychology, 74, 1464 –1480.
Greenwald, A. G., & Nosek, B. A. (2001). Health of the Implicit Associ-
ation Test at age 3. Zeitschrift fur Experimentelle Psychologie, 48,
85–93.
Karpinski, A., & Hilton, J. L. (2001). Attitudes and the Implicit Associa-
tion Test. Journal of Personality and Social Psychology, 81, 774 –788.
Kraiger, K., & Ford, J. K. (1985). A meta-analysis of race effects in
performance appraisal. Journal of Applied Psychology, 70, 56 65.
Kreft, I., & De Leeuw, J. (1998). Introducing multilevel modeling. Thou-
sand Oaks, CA: Sage.
Lykken, D. T. (1968). Statistical significance in psychological research.
Psychological Bulletin, 70, 151–159.
Maass, A., Castelli, L., & Arcuri, L. (2000). Measuring prejudice: Implicit
versus explicit techniques. In D. Capozza & R. Brown (Eds.), Social
identity processes: Trends in theory and research (pp. 96 –116). London:
Sage.
McConahay, J. B. (1986). Modern racism, ambivalence, and the Modern
Racism Scale. In J. F. Dovidio & S. L. Gaertner (Eds.), Prejudice,
discrimination, and racism (pp. 91–125). Orlando, FL: Academic Press.
McConahay, J. B., Hardee, B. B., & Batts, V. (1981). Has racism declined?
It depends upon who’s asking and what is asked. Journal of Conflict
Resolutions, 25, 563–579.
McConnell, A. R., & Leibold, J. M. (2001). Relations between the implicit
association test, explicit racial attitudes, and discriminatory behavior.
Journal of Experimental Social Psychology, 37, 435– 442.
Nisbett, R. E., & Wilson, T. D. (1977). Telling more than we can know:
Verbal reports on mental processes. Psychological Review, 84, 231–259.
Nunnally, J. C., & Bernstein, I. H. (1994). Psychometric theory (3rd ed.).
New York: McGraw-Hill.
Oppler, S. H., Campbell, J. P., Pulakos, E. D., & Borman, W. C. (1992).
Three approaches to the investigation of subgroup bias in performance
measurement: Review, results, and conclusions. Journal of Applied
Psychology, 77, 201–217.
Phelps, E. A., O’Connor, K. J., Cunningham, W. A., Gatenby, J. C.,
Funayama, E. S., Gore, J. C., & Banaji, M. R. (2000). Amygdala
activation predicts performance on indirect measures of racial bias.
Journal of Cognitive Neuroscience, 12, 729–738.
Plant, E. A., & Devine, P. G. (1998). Internal and external motivation to
respond without prejudice. Journal of Personality and Social Psychol-
ogy, 75, 811– 832.
Roberson, L., & Block, C. J. (2001). Racioethnicity and job performance:
A review and critique of theoretical perspectives on the causes of group
differences. Research in Organizational Behavior, 23, 247–325.
Roth, P. L., Huffcutt, A. I., & Bobko, P. (2003). Ethnic group differences
in measures of job performance: A new meta-analysis. Journal of
Applied Psychology, 88, 694 –706.
Rudman, L. A., & Kilianski, S. E. (2000). Implicit and explicit attitudes
toward female authority. Personality and Social Psychology Bulletin, 26,
1315–1328.
Schneider, B. (1972). Organizational climate: Individual preferences and
organizational realities. Journal of Applied Psychology, 56, 211–217.
Received November 19, 2002
Revision received December 18, 2003
Accepted December 22, 2003
562
RESEARCH REPORTS
This document is copyrighted by the American Psychological Association or one of its allied publishers.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
... Brescoll, 2011;Motro & Ellis, 2017;Rosette et al., 2008), in a location that is not disclosed in the manuscript (e.g. Johnson, Murphy, Zewdie, & Reichard, 2008;Kaufmann, Krings, & Sczesny, 2016;Ziegert & Hanges, 2005), or near the end of subjects' participation in a study, although even in these cases, there can be ambiguity regarding whether or not participants can alter questionnaire response after completing the manipulation check (e.g. Biernat & Hernandez et al., 2016;Rosette, Leonardelli, & Phillips, 2008). ...
... Organizational diversity research derived from experimental studies frequently utilizes student participants (e.g. Johnson et al., 2008;Livingston & Pearce, 2009;Rosette et al., 2008;Zapata et al., 2016;Ziegert & Hanges, 2005). Although the use of student samples in a laboratory setting provides researchers with some benefits, such as tightly controlled conditions where treatment effects can be more effectively isolated, it may also reduce the generalizability and replicability of findings (Singleton Jr. & Straits, 2005). ...
Article
This research is a pre-registered replication of Rosette, Leonardelli, and Phillips' (2008) seminal work in leadership categorization theory. Their work established race as a component to the business leader prototype and found evidence that when a leader was given credit for successful organizational performance, White leaders were evaluated more favorably than non-White leaders. As leadership exemplars are evolving, however, a need to reexamine these relationships has emerged. Results from our replications of their first and third studies showed minimal support for the argument that being White is a component of the business leader prototype. Additionally, across six separate studies, we found no conditions in which White leaders received more favorable evaluations than their non-White counterparts. Contrary to our expectations, we found that non-White leaders received marginally more favorable ratings than White leaders in four of our studies.
... That is, we test whether racial microaggressions may be a linking mechanism between racial bias and employee perceptions of abusive supervision. Because bias-and stereotype-motivated behavior have moved from being overt, direct and hostile, to more subtle, ambiguous and sometimes unintentional (Ziegert & Hanges, 2005), it is now critical to understand if manager's implicit bias could possibly predict these small acts of discriminatory behaviour that contribute to subordinates' perceptions of abusive supervision at work. Tepper et al. (2011) found that managers may view some employees as worthy of just treatment, that is, within their scope of justice, but other employees as "expendable, undeserving, exploitable, and irrelevant" (Opotow & Weiss, 2000: 478), that is justifiable targets of abusive supervision. ...
Article
Full-text available
We explore whether abusive supervision may occur more in mixed-race supervisor-subordinate dyads. Specifically, our model tests whether, in mixed-race dyads, a manager’s implicit racial bias may be associated with racial microaggressions, and, subsequently, subordinates’ perceptions of the degree to which that manager is an abusive supervisor. Social identity theory supports the why of these predictions. We also test when it may be possible for some managers to overcome their racial biases—by engaging in behaviors reflective of viewing their subordinates as individuals, rather than members of another race, via individuation theory. In this vein, we investigate a way in which race-based mistreatment and abusive supervision may be mitigated. We tested our predictions in 137 manager-employee dyads in two chemical manufacturing firms in South Africa. We found a positive relationship between manager implicit racial bias and abusive supervision, and that this relationship is lessened by individualized consideration–a moderator of the mediated effect of manager racial microaggressions on bias and abuse. Thus, our hypotheses were supported. We conclude with implications for victimized employees, and possible strategies to combat race-based aggression for organizations.
... These attitudes can impact workplace decisions related to recruitment, compensation and advancement opportunities (Dennis and Thomas, 2007). In fact, it has been argued that implicit attitudes more strongly predict discriminatory decisions than do explicit attitudes (Rudman and Glick, 2001;Ziegert and Hanges, 2005;Rooth, 2010). ...
Article
Full-text available
Mindfulness is becoming increasingly popular in the workplace. This likely relates to a growing body of research linking mindfulness to a range of psychological outcomes such as reduced anxiety, depression and increased subjective wellbeing. However, while mindfulness has received a great deal of attention in clinical research, the evidence for workplace relevant benefits is less established. Additionally, outside of clinical research, mindfulness studies have rarely been replicated. Recent evidence suggests that the cognitive skills cultivated during meditation may be instrumental in reducing biased thinking and increasing prosocial behaviour, but these findings have not been previously tested in a workplace setting. Specifically, mindfulness has been linked to reductions in implicit age bias, sunk-cost decision-making bias and increases in organisational citizenship behaviours (OCB). In two experiments using a workplace and laboratory sample, the present investigation aimed to test the reliability and generalisability of previous findings that a brief mindfulness meditation can reduce age and sunk-cost decision-making biases. To more directly test the potential positive benefits of mindfulness in a workplace setting, this study also investigated the impact of a mindfulness intervention on intention to perform OCB. While meditation significantly increased OCB intent, predictions relating to bias were not supported. Considerations for the degree to which empirical evidence aligns with claims in popular culture, along with implications for the practical uses of mindfulness in the workplace are explored.
... One way context may matter is via the "racial climate" of an institution, which may dictate the extent to which stereotyping impacts the job prospects of White versus non-White IPs. For example, research on employment discrimination outside of prison suggests that racial climate influences the impact of racial bias (e.g., Ziegert & Hanges, 2005). One way to test this possibility is to examine the racial heterogeneity of staff in a facility. ...
Article
Full-text available
This study assesses whether racial and ethnic disparities exist in prison industry employment and whether seemingly race‐ and ethnicity‐neutral eligibility requirements contribute to any such disparities. We examine whether there are racial/ethnic disparities in industrial prison work, the extent to which disparities are explained by administrative policies, and the conditions under which disparities are most pronounced. Using 10 years of prison administrative data from Ohio, this study employs multilevel and mediation analyses to examine the effects of race and ethnicity on the odds of working an industrial prison job. Results suggest that Black and Hispanic incarcerated persons (IPs) are less likely to work industry jobs than White IPs. The majority of this disparity stems from program requirements; however, some disparities maintain even when accounting for requirements. Black IPs who do not meet program requirements are less likely to work than White IPs who do not meet program requirements. Racial disparities are smaller in facilities with greater racial heterogeneity among correctional staff. In our discussion, we underscore how prison policies can contribute to racially and ethnically disparate incarceration experiences. The results suggest the importance of evaluating prison and other correctional policies that utilize selection criteria that appear race neutral but are likely to be disparate in their consequences. Moreover, policies aimed at diversifying staff may contribute to more equitable prison experiences for non‐White incarcerated people, although doing so does not directly address underlying policy problems that lead to inequalities.
Article
The algorithmic accountability literature to date has primarily focused on procedural tools to govern automated decision-making systems. That prescriptive literature elides a fundamentally empirical question: whether and under what circumstances, if any, is the use of algorithmic systems to make public policy decisions perceived as legitimate? The present study begins to answer this question. Using factorial vignette survey methodology, we explore the relative importance of the type of decision, the procedural governance, the input data used, and outcome errors on perceptions of the legitimacy of algorithmic public policy decisions as compared to similar human decisions. Among other findings, we find that the type of decision—low importance versus high importance—impacts the perceived legitimacy of automated decisions. We find that human governance of algorithmic systems (aka human-in-the-loop) increases perceptions of the legitimacy of algorithmic decision-making systems, even when those decisions are likely to result in significant errors. Notably, we also find the penalty to perceived legitimacy is greater when human decision-makers make mistakes than when algorithmic systems make the same errors. The positive impact on perceived legitimacy from governance—such as human-in-the-loop—is greatest for highly pivotal decisions such as parole, policing, and healthcare. After discussing the study’s limitations, we outline avenues for future research.
Article
Full-text available
The paper examines leader-unconscious bias and task accomplishment in manufacturing business organizations in Ogun state, Nigeria.
Article
Psychotic disorders are among the most highly stigmatized mental disorders, and individuals with psychosis experience significant exclusion from the community. Stigma reduction programs have done little to reduce social exclusion of individuals with psychosis, and there are significant limitations to the traditional stigma model as it applies to social exclusion. Herein, we present the Interactional Processing Model (IPM) of social exclusion towards individuals with psychosis. The IPM considers social exclusion to be the result of two interacting pathways with additional consideration for a feedback loop through which social exclusion sets in motion natural behavioural responses of individuals with psychosis that inadvertently perpetuates exclusion. The IPM considers initial social exclusion to be the result of an interaction between these two pathways. The first path aligns with the traditional stigma model and consists of the community becoming aware that an individual is diagnosed with a psychotic disorder and then excluding the individual based on pre-existing, generalized knowledge about the disorder. The second path to exclusion involves the observation of atypical behaviours from the individual, and generation of an individualized exclusion response. We provide initial empirical support for the IPM of social exclusion, outline testable hypotheses stemming from the model, and discuss implications for novel ways to consider both societal stigma reduction and personalized intervention.
Article
We evaluate the causes of the wage gap at the intersection of race, ethnicity and gender over time in the United States. We analyse the wage gaps for women of colour along three dimensions: relative to White women, relative to men of their respective race/ethnicity, and relative to White men. Using the American Community Survey, we replicate earlier findings based on the Current Population Survey data which show that, on average, Black women face an unexplained wage gap relative to White men that goes beyond the simple addition of the separate unexplained gender and racial wage gaps. This can be seen persistently between 1980 and 2019, and we find it is true across the entire wage distribution but especially notable at higher centiles. From 1990 through 2019, Black and Hispanic women saw stalled progress, while White women continued to make steady progress closing the wage gap relative to White men.
Article
Full-text available
The present research examines the assertion that individuals differ in the extent to which they seek to control the expression of prejudice. The authors developed the Motivation to Control Prejudiced Reactions Scale to assess this individual difference. Psychometric properties of the scale are reported, including its stable two-factor structure across samples. In addition, evidence regarding predictive validity is presented. The expression of racial prejudice on self-report measures was moderated by the extent to which respondents reported being motivated to inhibit prejudiced reactions. Specifically, the authors observed interactions between unobtrusive estimates of racial attitudes based on automatic attitude activation and scores on the Motivation to Control Prejudiced Reactions Scale when predicting self-reported evaluations. Motivated individuals expressed less prejudiced responses even if their unobtrusive estimates revealed automatically activated negativity in response to Blacks. In contrast, the less motivated provided self-reports consistent with their automatically activated attitudes.
Article
Full-text available
Attitudes toward female authority and their relationship to gender beliefs were examined using implicit and explicit measures of each. Implicit attitudes covaried with implicit gender authority beliefs (i.e., linking men to high-authority and women to low-authority roles). Explicit attitudes covaried with explicit gender authority beliefs, feminist identification, and hostile sexism. Thus, gender authority beliefs may influence both conscious and unconscious prejudice against female authorities. Although women showed less explicit prejudice than did men, their implicit attitudes were similarly negative. Finally, the relationship found between two different response latency methods (a priming task for attitudes, a categorization task for beliefs) supports the assumption that implicit measures assess similar constructs (i.e., automatic associations in long-term memory).
Article
Three experiments provided evidence that intergroup bias occurs automatically under minimal conditions, using the Implicit Association Test (IAT; A. G. Greenwald, D. E. McGhee, & J. L. K. Schwartz, 1998). In Experiment 1, participants more readily paired in-group names with pleasant words and out-group names with unpleasant words, even when they were experienced only with the in-group and had no preconceptions about the out-group. Participants in Experiment 2 likewise showed an automatic bias favoring the in-group, even when in-group/out-group exemplars were completely unfamiliar and identifiable only with the use of a heuristic. In Experiment 3, participants displayed a pro-in-group IAT bias following a minimal group manipulation. Taken together, the results demonstrate the ease with which intergroup bias emerges even in unlikely conditions.
Article
Evidence is reviewed which suggests that there may be little or no direct introspective access to higher order cognitive processes. Subjects are sometimes (a) unaware of the existence of a stimulus that importantly influenced a response, (b) unaware of the existence of the response, and (c) unaware that the stimulus has affected the response. It is proposed that when people attempt to report on their cognitive processes, that is, on the processes mediating the effects of a stimulus on a response, they do not do so on the basis of any true introspection. Instead, their reports are based on a priori, implicit causal theories, or judgments about the extent to which a particular stimulus is a plausible cause of a given response. This suggests that though people may not be able to observe directly their cognitive processes, they will sometimes be able to report accurately about them. Accurate reports will occur when influential stimuli are salient and are plausible causes of the responses they produce, and will not occur when stimuli are not salient or are not plausible causes.
Article
This special issue of the Journal of Personality and Social Psychology: Attitudes and Social Cognition addresses issues of the measurement and the malleability of implicit prejudice and stereotypes. The findings raise fundamental questions about the assumptions underlying the assessment of implicit prejudice, particularly with regard to the widely used Implicit Association Test (A. Greenwald, D. McGhee, & J. Schwartz, 1998) and the assumption of extant models of prejudice and stereotyping that implicit biases are automatically and invariantly activated when perceivers come in contact with members of stigmatized groups. Several of the articles show that contextual manipulations produce reductions in implicit manifestations of prejudice and stereotyping. The articles in this issue, in challenging conventional wisdom, are thought provoking and should be generative in the field's ongoing efforts to understand the role of implicit (and explicit) processes involved in prejudice and stereotyping. (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Article
Three methods of assessing subgroup bias in performance measurement commonly found in the literature are identified. After a review of these approaches, findings are reported from analyses of data collected in the United States Army's Project A (J. P. Campbell, 1987). Correlations between nonrating performance measures and supervisor ratings were generally not moderated by race, but correlations between nonrating indicators of negative performance and ratings assigned by peers were. In addition, significant interactions between rater and ratee race on performance ratings were not eliminated when variance in the nonrating measures was removed from the ratings provided by Black and White raters. Conclusions about the magnitude and nature of bias in supervisor and peer ratings are discussed.
Article
What was noted by E. J. Langer (1978) remains true today; that much of contemporary psychological research is based on the assumption that people are consciously and systematically processing incoming information in order to construe and interpret their world and to plan and engage in courses of action. As did E. J. Langer, the authors question this assumption. First, they review evidence that the ability to exercise such conscious, intentional control is actually quite limited, so that most of moment-to-moment psychological life must occur through nonconscious means if it is to occur at all. The authors then describe the different possible mechanisms that produce automatic, environmental control over these various phenomena and review evidence establishing both the existence of these mechanisms as well as their consequences for judgments, emotions, and behavior. Three major forms of automatic self-regulation are identified: an automatic effect of perception on action, automatic goal pursuit, and a continual automatic evaluation of one's experience. From the accumulating evidence, the authors conclude that these various nonconscious mental systems perform the lion's share of the self-regulatory burden, beneficently keeping the individual grounded in his or her current environment.
Article
In recent years as public opinion polls have shown a decline in racist responses, white Americans have strongly resisted school desegregation and affirmative action programs. Hence, there has been a debate over the extent to which racism has really declined. The theory of modern racism addresses these issues, distinguishing between old-fashioned racial beliefs recognized by everyone as racism and a new set of beliefs arising from the conflicts of the civil rights movement. The theory proposes that antiblack feeling remains high and has been displaced from the socially undesirable old-fashioned beliefs onto the new beliefs where the racism is not recognized. Three experiments were performed; results showed that, regardless of context, the old-fashioned items were perceived as more likely to reveal prejudice. The results are discussed in terms of their significance for opinion polling and continuing racial conflict in America.