Organizational Justice and Black Applicants' Reactions to Affirmative Action

Abstract

Using organizational justice as a guiding framework, the authors studied perceptions of affirmative action programs by presumed beneficiaries. Three conceptual issues were addressed: (a) the content of different affirmative action plans; (b) the 3-way interaction among distributive, procedural, and interactional justice; and (c) the distinction between outcome favorability and distributive justice. These ideas were tested with a sample of Black engineering students who responded to 1 of 6 plans. Participants distinguished among the various plans, with some policies being viewed as more fair than others. In addition, a 3-way interaction among the 3 types of organizational justice was observed. Specifically, the 2-way interaction between distributive and interactional fairness was only significant when procedural justice was low. Implications for organizational justice and for the design of affirmative action programs are discussed.

Full text

Organizational Justice and Black Applicants’ Reactions
to Affirmative Action
Russell Cropanzano and Jerel E. Slaughter
University of Arizona
Peter D. Bachiochi
Eastern Connecticut State University
Using organizational justice as a guiding framework, the authors studied perceptions of affirmative action
programs by presumed beneficiaries. Three conceptual issues were addressed: (a) the content of different
affirmative action plans; (b) the 3-way interaction among distributive, procedural, and interactional
justice; and (c) the distinction between outcome favorability and distributive justice. These ideas were
tested with a sample of Black engineering students who responded to 1 of 6 plans. Participants
distinguished among the various plans, with some policies being viewed as more fair than others. In
addition, a 3-way interaction among the 3 types of organizational justice was observed. Specifically, the
2-way interaction between distributive and interactional fairness was only significant when procedural
justice was low. Implications for organizational justice and for the design of affirmative action programs
are discussed.
Keywords: affirmative action, organizational justice, diversity
Variously understood (and often misunderstood) affirmative
action (AA) is among the most controversial personnel procedures
facing work organizations. President Lyndon Johnson originally
created AA in 1965 pursuant to Executive Order 11246 (Skrentny,
1996). The relevant definition can be found in Subpart B, Section
202(1):
The contractor will take affirmation action to ensure that applicants
are employed, and that employees are treated during employment,
without regard to their race, color, religion, sex or national origin.
Such action shall include, but not be limited to the following: em-
ployment upgrading, demotion, or transfer; recruitment or recruitment
advertising; layoff or termination; rates of pay or other forms of
compensation; and selection for training, including apprenticeship.
In other words, AA involves specific organizational policies to
recruit, promote, train, select, and/or retain workers who are mem-
bers of protected classes (Deblieux, 1996; Sovereign, 1994). This
brief definition captures some of the tension surrounding AA. On
the one hand (per Executive Order 11246), AA should treat people
“without regard to their race, religion, sex or national origin.” On
the other hand, given that societal disparities already exist, affir-
mative steps (such as recruitment, training, etc.) are recommended
strategies for achieving the ultimate goal of equality. Given these
dual considerations, AA is defined as a partial remedy to the
pernicious effects of discrimination (Cross, 1994; Skrentny, 1996),
while also being criticized for unfairly disadvantaging majority
workers (Mosley & Capaldi, 1996; Sowell, 1990).
As the means by which AA is pursued can appear to contradict
the goal of equal treatment (at least in the opinion of some people),
issues of social justice loom at the heart of the debate (cf. Bobo &
Kluegel, 1993; Bobocel, Son Hing, Davey, Stanley, & Zanna,
1998; Kravitz et al., 1997; Velasquez, 1992). Fairness has pro-
vided a useful conceptual lens through which to view AA (e.g.,
Bobocel, Davey, Son Hing, & Zanna, 2001; Nacoste, 1987; Leck,
Saunders, & Charbonneau, 1996; Singer, 1992, 1993, 1996). Gen-
erally speaking, research has found that affirmative action plans
(AAPs) affect perceptions of distributive justice (the fairness of the
outcomes), procedural justice (the fairness of the allocation pro-
cess), and interactional justice (the fairness of interpersonal treat-
ment). As we discuss below, it might be helpful to reframe some
of the questions previously asked about the perceived fairness of
AA. Four issues concern us here.
First, we were especially interested in examining conceptual and
practical issues related to AA among Blacks. Historically, Blacks
have been less likely to serve as research participants than their
White counterparts (e.g., Graham, 1992; James, 1997). This is
problematic for research on organizational justice, where Whites
and Blacks sometimes have different perceptions (e.g., Davidson
& Friedman, 1998; Friedman & Davidson, 1999). Therefore, we
sought to examine perceptions of AA justice among presumed
beneficiaries.
Second, the public often uses the term affirmative action in a
somewhat loose manner. AA is a collection of programs that are
potentially available to organizations. Individuals may have dif-
fering judgments as to the fairness or unfairness of each program.
For example, one might strongly favor enhanced recruitment while
still opposing preferential hiring policies. For this reason, blanket
judgments of “affirmative action” are often less meaningful than
evaluations of specific programs (for evidence, see Kravitz, 1995;
Kravitz & Platania, 1993).
Russell Cropanzano and Jerel E. Slaughter, Department of Management
and Policy, Eller College of Management, University of Arizona; Peter
D. Bachiochi, Department of Psychology, Eastern Connecticut State
University.
Correspondence concerning this article should be addressed to Russell
Cropanzano, Department of Management and Policy, Eller College of
Management, University of Arizona, McClelland Hall, Room 405, P.O.
Box 210108, Tucson, AZ 85721. E-mail: russell@eller.arizona.edu
Journal of Applied Psychology Copyright 2005 by the American Psychological Association
2005, Vol. 90, No. 6, 1168–1184 0021-9010/05/$12.00 DOI: 10.1037/0021-9010.90.6.1168
1168

Third, popular discourse has often treated the fairness of AA
broadly, without considering that justice has at least three facets—
distributive, procedural, and interactional (e.g., Slaughter, Sinar, &
Bachiochi, 2002). Not only can an AAP impact each type of
justice, but the three justice types also might further interact to
affect additional criteria such as attraction to the organization as a
place to work (Goldman, 2003; Skarlicki & Folger, 1997). There-
fore, research emphasizing only main effects of justice lacks
valuable information that can only be obtained by considering
moderator effects. For this reason it is important to consider how
AA affects each of the three major types of justice, as well as how
these three types of justice interact to predict responses. We return
to this issue in a moment.
Fourth, greater attention to the conceptualization of outcomes as
they relate to AA is needed. It is especially important to separate
outcome favorability from outcome fairness (e.g., distributive jus-
tice). Both fairness and favorability are important parts of the
national discourse on AA (see Cahn, 2002; Cross, 1994; Curry &
West, 1996; Mosley & Capaldi, 1996). However, they are not
conceptually identical. A favorable outcome is one that is consis-
tent with an individual’s personal interests or desires. A fair
outcome is one that is consistent with moral standards or norma-
tive principles. For example, Deutsch (1975, 1985) suggested that
there are three prototypical allocation standards: equity (to each
according to his or her performance), equality (to each the same),
and need (to each according to his or her needs).
As an illustration, consider the situation facing a job seeker who
is rejected for a position. Suppose further that this individual
discovers that he or she was less qualified than was the person who
received the position. Therefore, the decision was fair, because the
job was assigned in conformance with acceptable standards of
conduct. However, the rejection was unfavorable, because the
applicant did not receive the position he or she wanted. Research
suggests that fairness and favorability are different constructs with
distinct nomological nets (Skitka, Winquist, & Hutchinson, 2003;
Van den Bos, 1998; Van den Bos, Wilke, Lind, & Vermunt, 1998).
The adoption of an AA program could make outcomes more or
less favorable to some, but this is not the same thing as making
them more or less fair. Thus, researchers’ understanding could
benefit by distinguishing outcomes that are unfair from those that
simply are not beneficial. In the present study we consider distrib-
utive injustice beyond the effects of unfavorability. By paying
close attention to each of these concerns, our hope is to frame
questions involving AA fairness in a manner helpful to researchers
and practitioners alike.
Effects of AA on Perceptions of Different Justice Types
Although various authors have examined justice perceptions of
AA, much of this research has tended to study global perceptions.
In the present investigation, we study distributive, procedural, and
interactional justice. The available research, although limited, sug-
gests that this could be a promising strategy. For example, in a
survey of Canadian workers, Leck et al. (1996) determined that
various aspects of AA programs can lead employees to report less
procedural and distributive justice. These feelings of injustice, in
turn, can further impact how a female or minority employee is
treated by his or her new coworkers. Less research has addressed
the impact of interactional justice on applicant responses to AA.
However, it seems to be important. In three role-playing experi-
ments, Singer (1993, chap. 6) found that when an appropriate
justification was provided for preferential treatment, individuals of
two different ethnicities (European and Asian) responded more
positively to an AA procedure. Provision of reasonable justifica-
tion is, of course, an important part of interactional justice (Bies,
1987).
Perceptions of Different Forms of AA
Especially within the popular press, there has been a great deal
of misunderstanding regarding AA (for a historical review, see
Skrentny, 1996). Some individuals view AA as simply preferential
policies. In fact, AA actually refers to a multitude of programs that
promote greater representation of underrepresented groups (Krav-
itz et al., 1997). We propose that these different types of AAPs are
likely to evoke different levels of justice perceptions.
Plans Studied in the Present Investigation
The plans studied in the present investigation are presented in
Table 1. In choosing the plans for inclusion, we sought to include
plans on which additional research is needed. We also sought to
include plans that would provide enough variance on the outcome
variables to test the justice-related hypotheses. The plans used in
the present study had all been used previously in at least one of
three published studies (Kravitz, 1995; Kravitz & Klineberg, 2000;
or Slaughter et al., 2002).
As Table 1 shows, the Control Plan indicates that the goal of the
organization is to reverse past discrimination without the use of
AA, and managers are told that hiring decisions are to be based
entirely on past experience and test scores. The Eliminate Discrim-
ination Plan specified that managers who discriminate will be
evaluated poorly and will be fired. The stipulations included in the
Eliminate Discrimination Plan were included in the remaining
plans, and the remaining plans added one of four stipulations
beyond those provided by the Eliminate Discrimination Plan. The
Recruitment Plan added that the AAP involved the recruitment of
Black applicants. The Training Plan added that a training program
teaching interviewing and job choice skills was available for Black
applicants. The Tiebreak Plan added that a Black applicant would
be hired over a White applicant if the applicants were equally
qualified. The Preferential Treatment Plan added that preferential
treatment is given to Black applicants, in that a Black applicant
could be hired over a more qualified White applicant.
Note that the Control, Eliminate Discrimination, and Recruit-
ment Plans could be considered race-blind plans because the
stipulations of these policies are such that race cannot directly
affect the hiring decision. The Tiebreak and Preferential Plans
could be considered race-conscious plans because race is highly
likely to affect hiring decisions with those plans in place. In
addition, although the Training Plan stipulates that hiring decisions
are made without regard to the applicant’s race (race blind), the
plan also stipulates a special training program for Black applicants,
which ostensibly would increase their chances of receiving an
offer (race conscious). Thus, it was placed in the race-conscious
category. When predicting the perceived fairness of the plans,
therefore, we categorized the plans as either race blind or race
conscious.
1169
JUSTICE AND AFFIRMATIVE ACTION

Effects of Plan Content Manipulation on Fairness
Perceptions
Although conventional wisdom may argue against it, research
shows that recipients of AA often react negatively to race-
conscious AAPs (Singer, 1993, chap. 5). Those plans that are
designed to provide the strongest benefits to Black applicants and
employees are often seen as having less distributive and procedural
justice than are those plans with more modest benefits (e.g.,
Slaughter, Bulger, & Bachiochi, in press; Slaughter et al., 2002).
Below, we review the previous theory and research that led to the
study’s hypotheses.
Under some circumstances, members of a group that should
benefit from an AA program can be inadvertently harmed. For
example, Heilman and colleagues have illustrated the potential
stigma of gender-based preferential selection. Heilman, Block, and
Lucas (1992) found that a gender-based AA label was linked to
lower perceptions of competence. Beneficiaries of gender-based
preferential selection not only had more negative perceptions of
their own performance and leadership ability (Heilman, Simon, &
Repper, 1987), but they also believed others had lower expecta-
tions of their competence (Heilman & Alcott, 2001). These neg-
ative self-perceptions also may have an unanticipated outcome.
When women thought they were AA beneficiaries, they provided
lower performance evaluations and competence ratings of other
women, but not men (Heilman, Kaplow, Amato, & Stathatos,
1993). These findings suggest that AA may carry a threat to those
it was intended to help.
Consistent with the work of Heilman and her colleagues, Turner
and Pratkanis (1994) have conceived of AA as help toward in-
tended beneficiaries. The authors suggested that AA can be ben-
eficial only if the specific language and nature of the plan carry
messages of self-support, such as when they highlight the helper’s
caring for the recipient and do not imply the recipient’s failure or
inferiority in needing help. Other AAPs might instead carry a
message of self-threat. For example, a threatening plan might
imply that someone lacks the requisite qualifications or could not
attain a job without assistance. When such a threat exists, it is
likely that the AAP will be perceived as harmful. When one
considers the language and the actions specified in each of the
AAPs presented to the respondents in this study (see Table 1), it is
Table 1
Affirmative Action Plans
Plan Description
Control Sandeman strives to reverse past discrimination suffered by Black employees in the organization without the use of
affirmative action. Managers are told that hiring decisions are to be based entirely on the applicant’s past experience and
scores on a set of selection tests. Thus, hiring decisions are made without regard to the applicant’s race.
Eliminate Discrimination The goal of the affirmative action plan employed by Sandeman is to reverse past discrimination suffered by Black
employees in the organization. Their affirmative action plan involves the complete elimination of discrimination against
Blacks. It is emphasized to managers that they must not discriminate against Blacks when making hiring decisions and
that any manager who discriminates will receive poor performance appraisals and may be fired. Hiring decisions are to
be based entirely on the applicant’s past experience and scores on a set of selection tests. Thus, hiring decisions are
made without regard to the applicant’s race.
Recruiting The goal of the affirmative action plan employed by Sandeman is to reverse past discrimination suffered by Black
employees in the organization. Their affirmative action plan involves the complete elimination of discrimination against
Blacks and the active recruitment of Black applicants. It is emphasized to managers that they must not discriminate
against Blacks when making hiring decisions and that any manager who discriminates will receive poor performance
appraisals and may be fired. Hiring decisions are to be based entirely on the applicant’s past experience and scores on a
set of selection tests. Hiring decisions are made without regard to the applicant’s race.
Training The goal of the affirmative action plan employed by Sandeman is to reverse past discrimination suffered by Black
employees in the organization. Their affirmative action plan involves the complete elimination of discrimination against
Blacks and the operation of a special training program for Black applicants. It is emphasized to managers that they must
not discriminate against Blacks when making hiring decisions and that any manager who discriminates will receive poor
performance appraisals and may be fired. Hiring decisions are to be based entirely on the applicant’s past experience
and scores on a set of selection tests. In addition, a training program for possible Black applicants is used. Blacks
accepted to the training program are taught basic interviewing and job-choice skills. After completion of the training
program, they are expected to apply for a position. Hiring decisions are made without regard to the applicant’s race.
Tiebreak The goal of the affirmative action plan employed by Sandeman is to reverse past discrimination suffered by Black
employees in the organization. Their affirmative action plan involves the complete elimination of discrimination against
Blacks and attention to the proportion of qualified Blacks who are hired. It is emphasized to managers that they must
not discriminate against Blacks when making hiring decisions and that any manager who discriminates will receive poor
performance appraisals and may be fired. Hiring decisions are to be based primarily on the applicant’s past experience
and scores on a set of selection tests. However, if a Black applicant and a White applicant are equally qualified, the
Black applicant is to be hired.
Preferential Treatment The goal of the affirmative action plan employed by Sandeman is to reverse past discrimination suffered by Black
employees in the organization. Their affirmative action plan involves the complete elimination of discrimination against
Blacks and attention to the proportion of qualified Blacks who are hired. It is emphasized to managers that they must
not discriminate against Blacks when making hiring decisions and that any manager who discriminates will receive poor
performance appraisals and may be fired. Thus, hiring decisions are to be based primarily on the applicant’s past
experience and scores on a set of selection tests. In addition, strong preferential treatment is given to Black applicants.
In some cases, Black applicants can be hired even if they have somewhat less experience and poorer test scores than
White applicants.
1170
CROPANZANO, SLAUGHTER, AND BACHIOCHI

clear that race-conscious plans are more likely to carry a message
of self-threat. For example, the Tiebreak AAP specifies that hiring
preference will be given to Black applicants over White applicants
if they are equally qualified, which could potentially lead to a
Black applicant being hired to the exclusion of an equally qualified
White applicant. The Preferential Treatment AAP goes even be-
yond this stipulation to specify that preference may be given to
Black applicants even if they are less qualified than White appli-
cants. Even the Training AAP could be considered self-
threatening, because it implies that Black applicants lack basic
interviewing skills and that they might not be selected without
additional training.
On the basis of previous theory and the research discussed
above, we expected that the race-conscious plans would lead to
more negative perceptions of perceived justice. In addition to our
overall expectations that these types of AA will engender lower
general fairness reactions, theory and research suggest that plan
type will affect the three specific justice types. For example,
distributive justice concerns the fairness of the outcomes that are
distributed in a given situation (Adams, 1965). High levels of
distributive justice are generally perceived when outcomes are
distributed on the basis of well-defined rules, such as equity or
equality (Colquitt, 2001; Cropanzano & Greenberg, 1997). If we
consider the differences between the race-conscious plans (Train-
ing, Tiebreak, and Preferential Treatment) and race-blind plans
(Control, Eliminate Discrimination, and Recruitment) in terms of
the extent to which the plans follow each of the rules, it becomes
clear that the race-blind plans attempt to distribute outcomes in an
equitable manner. These plans specifically indicate that hiring
decisions are to be based entirely on the applicant’s past perfor-
mance and selection test scores. The remaining plans, however,
would likely violate the equity rule. The Tiebreak and Preferential
Treatment Plans are in clear violation of the equity rule, because
they suggest preference to Black applicants with equal or lower
qualification. The Training Plan, though less explicit than the
Tiebreak and Preferential Treatment Plans, also violates the equity
rule because Black interviewees would have had the opportunity to
receive special training, making it more likely that a Black appli-
cant would be hired over an otherwise equally qualified non-Black
applicant. Thus, we hypothesized the following:
Hypothesis 1a: The Training, Tiebreak, and Preferential
Treatment AAPs will be rated lower on distributive justice
than the Control, Eliminate Discrimination, and Recruitment
AAPs.
Procedural justice judgments tend to be somewhat more com-
plicated than distributive justice judgments, in part because of the
large number of rules that have been identified to mark a particular
procedure as just or unjust. Organizational justice scholars have
identified many such rules that may be more or less important,
depending on the context. For example, Leventhal (1976) identi-
fied six procedural justice rules. He proposed that procedures
could be considered to be fair to the extent that they are (a) applied
consistently, (b) free from bias, (c) accurate, (d) correctible, (e)
representative of all concerns, and (f) based on prevailing ethical
standards. Other researchers have studied procedural rules that
may apply in judging the fairness of specific procedures. Folger,
Konovsky, and Cropanzano (1992) specified three rules that might
be used to judge the fairness of the performance evaluation pro-
cess. Gilliland (1993) proposed that nine different rules can be
used to judge the procedural fairness of personnel selection sys-
tems. Of all of these rules, the one most relevant to AAP content
is that of consistency. The race-blind plans seek to apply rules in
a consistent manner to all applicants and employees, because all
applicants are treated similarly when they enter the applicant pool.
The Training, Tiebreak, and Preferential Treatment Plans either
implicitly (through special training) or explicitly (through prefer-
ential treatment) apply different treatment to Black and non-Black
applicants, violating the rule of consistency. Therefore, we hypoth-
esized the following:
Hypothesis 1b: The Training, Tiebreak, and Preferential
Treatment AAPs will be rated lower on procedural justice
than the Control, Eliminate Discrimination, and Recruitment
AAPs.
We also expected the race-blind and race-conscious AAPs to
differ in terms of perceived interactional justice. Interactional
justice is the fairness of interpersonal treatment as procedures are
enacted (Bies & Moag, 1986). Interactional justice perceptions are
more favorable to the extent that individuals are treated with
respect and sensitivity and are given reasonable explanations for
decisions (Colquitt, 2001). Although none of the AAPs provide
explanations for why the specific AAP is in place, it is reasonable
to predict that plans that specify race-blind procedures would be
evaluated as more justifiable than plans that specify race-conscious
procedures (Robinson, Seydel, & Douglas, 1998). Moreover, as far
as communicating respect for all groups, plans that specify training
programs and preferential treatment could be construed as disre-
spectful and stigmatizing, because they imply that Black applicants
would have difficulty securing job offers without those plans in
place (Slaughter et al., in press). Therefore, we hypothesized the
following:
Hypothesis 1c: The Training, Tiebreak, and Preferential
Treatment AAPs will be rated lower on interactional justice
than the Control, Eliminate Discrimination, and Recruitment
AAPs.
We also theorized that the degree to which the plans were
perceived as fair would be related to individuals’ overall percep-
tions of organizational attractiveness and their intentions to apply
for a position. Social inference theory (Tyler & Caine, 1981) and
theories of interdependence (Thibaut & Kelly, 1959) suggest that
individuals use perceived fairness of policies to make fairness
judgments of the entities that produce the policies. These judg-
ments, in turn, determine whether individuals will want to interact
with an entity in the future (Cropanzano, Byrne, Bobocel, & Rupp,
2001). Similarly, fairness heuristic theory (Lind, 2001) suggests
that because ceding authority to another person or to an organiza-
tion may lead to exploitation, people feel uneasy when they are
considering doing so. Therefore, individuals decide whether they
can trust the person or organization by using information encoun-
tered early in the relationship (van den Bos et al., 1998). As a
result, fairness judgments regarding an organization’s AAP are
likely to carry over to judgments of the organizational decision
makers (e.g., Slaughter et al., in press). At least two variables in
1171
JUSTICE AND AFFIRMATIVE ACTION

particular are critical to understanding applicant responses to se-
lection procedures: organizational attractiveness and intention to
apply. Available work suggests that these variables are predicted
by perceptions of workplace fairness (Cropanzano & Wright,
2003; Gilliland, 1993; Gilliland & Steiner, 2001; Judge & Bretz,
1992). Moreover, although theories of interdependence seem to
refer to overall perceptions of the fairness of the entity, this logic
can be extended naturally to distributive, procedural, and interac-
tional justice. Put simply, applicants should be more attracted to
organizations that implement policies that (a) distribute outcomes
fairly, (b) represent procedures designed to distribute outcomes in
a fair manner, and (c) indicate respect for the applicant. On the
basis of this theory and research, we hypothesized the following:
Hypothesis 2: Perceptions of (a) distributive justice, (b) pro-
cedural justice, and (c) interactional justice will be positively
related to organizational attractiveness.
Hypothesis 3: Perceptions of (a) distributive justice, (b) pro-
cedural justice, and (c) interactional justice will be positively
related to intentions to apply.
Three-Way Interaction Among the Three Types of Justice
We were also concerned with relationships among the three
types of organizational justice. In research on AA, scholars have
tended to treat distributive, procedural, and interactional justice
separately, not always giving sufficient emphasis to moderated
relationships (cf., Leck et al., 1996; Slaughter et al., 2002). While
useful, this main effect approach is limited because organizational
justice research suggests that the perceived attractiveness of an
organization, as well as intentions to apply, should be predicted by
the interaction among the three types of fairness. As an illustration
of this point, we briefly review the relevant literature.
In their studies of workplace fairness, Skarlicki and Folger
(1997) and Goldman (2003) have maintained that what you receive
(distributive justice) has stronger or weaker effects depending on
how you are treated (interactional justice) and on the allocation
process (procedural justice). Put differently, when forming opin-
ions of an organization or of an organizational decision maker,
individuals attend to at least three cues—the outcome, interper-
sonal treatment, and the allocation procedure. These cues are
combined multiplicatively in order to influence one’s evaluation.
In other words, different types of justice interact to predict various
work-relevant criteria (for reviews, see Brockner, 2002; Brockner
& Wiesenfeld, 1996; Cropanzano & Folger, 1991; Cropanzano &
Schminke, 2001).
Additionally, this interaction has a particular appearance. Skar-
licki and Folger (1997) and Goldman (2003) described its form in
terms of a qualified two-way interaction. As they put it, the
two-way interaction between distributive justice and interactional
justice is only significant when procedural justice is low. To
illustrate this effect, we can begin with the two-way interaction
between distributive and interactional justice. According to Brock-
ner and Wiesenfeld (1996), distributive justice predicts better
when interactional justice is low and worse when interactional
justice is high. Stated concretely, individuals tend to react less
negatively to an unfair distribution if they are treated with dignity
and respect. The three-way effect comes into being when this
two-way effect is qualified. In particular, when the procedure is
judged to be fair, then this can mitigate the impact of both an unfair
distribution and unfair interpersonal treatment. Therefore, just
procedures render the two-way interaction between distribution
and interaction nonsignificant. Of course, when procedural justice
is low, the two-way interaction should continue to be an effective
predictor. This is a three-way interaction because the two-way
interaction depends on the value of a third variable.
In support of these ideas, Skarlicki and Folger (1997) found that
the three-way interaction explained a significant amount of vari-
ance in organizational retaliation behaviors beyond the main ef-
fects and two-way interactions. Specifically, Skarlicki and Folger
observed that the two-way interaction of distributive and interac-
tional justice was only significant when procedural justice was
low. This three-way interaction was replicated in a recent paper by
Goldman (2003), who studied the propensity of terminated em-
ployees to file legal claims. As in Skarlicki and Folger’s (1997)
study, Goldman found that distributive and interactional justice
interacted to predict legal claiming. Once again, this two-way
interaction was only significant when procedural justice was low.
This research is important, but it has yet to be applied to the AA
or recruitment literatures. Skarlicki and Folger and Goldman im-
plied that ratings of organizational attractiveness and application
intentions might be impacted by the interaction among the three
types of fairness. That is, the multiplicative relationship among
distributive, procedural, and interactional justice could explain
additional variance beyond any additive relationships. On the basis
of this research, we offer the following hypothesis:
Hypothesis 4: There will be a three-way interaction among
distributive, interactional and procedural justice that predicts
(a) attractiveness of organizations using AA and (b) inten-
tions to apply for a job with the organization. In the case of
both of these criterion variables, the two-way interaction
between distributive justice and interactional justice will only
be significant when procedural justice is low.
Distinguishing Outcome Favorability From
Distributive Justice
Finally, we were concerned with the typical operationalization
of outcomes. There is a need to distinguish unfavorability from
distributive injustice. The former refers to whether an event is
beneficial to a particular respondent, whereas the latter refers to the
morality or appropriateness of a given allocation. Fairness, as
opposed to outcome favorability, refers to a violation of some
moral standard. In the case of distributive injustice, this would
refer to a situation where one received less than he or she earned,
deserved, or was otherwise entitled to.
Although judgments of distributive justice and outcome favor-
ability are correlated (Messick & Sentis, 1979, 1983), their rela-
tionship is sometimes modest (Cropanzano & Greenberg, 1997;
Tyler & Blader, 2000). Moreover, outcome fairness and favorabil-
ity have separate nomological networks (Skitka et al., 2003; Van
den Bos, 1998). Conflating an unjust outcome with a harmful one
creates interpretive problems because it is unclear which con-
cept—justice or economic interest—is driving observed relation-
ships. Given this concern, the present study will follow Bauer,
1172
CROPANZANO, SLAUGHTER, AND BACHIOCHI

Maertz, Dolen, and Campion (1998) and control for outcome
favorability in all of our analyses.
The distinction between injustice and unfavorability takes on
special meaning when testing the interaction among the different
types of fairness. Some authors have described the process by
outcome interaction as involving both distributive justice and also
outcome favorability (e.g., Brockner & Wiesenfeld, 1996; Cropan-
zano & Folger, 1991), whereas others have emphasized only
outcome favorability (e.g., Bauer et al., 1998; Brockner, 2002;
Cropanzano & Konovsky, 1996; Weiss, Suckow, & Cropanzano,
1999), and still others have focused exclusively on distributive
justice (e.g., Goldman, 2003; Skarlicki & Folger, 1997). Although
each of these models is defensible, it is important to rule out
plausible alternatives.
Following Goldman (2003) and Skarlicki and Folger (1997), the
theoretical model used here suggests that distributive justice
should interact with procedural and interactional justice. Conse-
quently, the analyses used here must serve two objectives. First,
they must specify that the predicted interaction (i.e., the one
involving distributive justice) is significant. Second, they must rule
out the potentially confounding impact of the unpredicted effects
(i.e., those involving outcome favorability).
To accomplish the first goal, we conducted moderated regres-
sion. Following from suggestions made by Aguinis (1995) we first
entered all of the main effects and two-way interactions into the
regression equation. Afterward, we assessed the incremental con-
tribution of the three-way interaction involving distributive justice.
To accomplish the second goal, we controlled for unfavorability in
all analyses that involve distributive justice, including the relevant
two-way and three-way interactions. Support for Hypothesis 4 can
be presumed to the extent that the Distributive Justice ⫻ Proce-
dural Justice ⫻ Interactional Justice term remains significant be-
yond the effect of these control variables.
Previous work testing the three-way interaction has not used all
of these control variables. Potentially, this could lead to rather
severe conceptual confusion, as one might mistakenly attribute to
distributive justice what should more appropriately be attributed to
outcome favorability. This issue is serious because perceived fair-
ness and favorability are correlated (Skitka et al., 2003). If favor-
ability is not controlled, it is difficult to make unambiguous infer-
ences about fairness given that variance allegedly accounted for by
justice could potentially be attributed to favorability. We might,
for example, conclude that the distributive fairness is more impor-
tant than it actually is. The problem is not only a psychometric one.
Rather, the matter is more substantive. To determine whether
distributive justice matters, it is important to examine its impact
beyond the effect of favorability.
Method
Participants
A national organization of Black engineering students, whose mission is
to increase the number of culturally responsible Black engineers, partici-
pated in the study. Members responded to a special research section of an
annual employer preference survey sponsored by the organization to yield
information for an annual “employer of choice” issue of their quarterly
magazine. Approximately 85% of the membership are engineering majors
representing every area of engineering, and the remainder are from various
science disciplines. Graduate students represent 10% of the membership
(the remaining 90% are undergraduate students), and female representation
in this group has recently approached 50%. Although engineering is
thought to be male dominated, women compose a fairly large percentage of
degree recipients, particularly among Blacks (National Science Founda-
tion, 2003). The survey was conducted online in the fall of 2002 and the
spring of 2003. An e-mail was sent to all members of the organization,
asking them to complete an online survey. Three hundred and forty-nine
members, representing approximately 5% of the organization’s overall
membership, responded. The sample was 52% male, with a mean age of
22.7 years, nearly identical to that of the organization as a whole.
AAPs
In the designated special research section of the online survey, respon-
dents read the following statement before they read the AAP:
We will be asking you several questions about the statement below, so
please read it carefully. Assume that you are searching for a job and
you find a company that fits your professional skills and interests. In
their brochure you see the following statement.
They were then asked to read one of six AAPs (no AA, Eliminate
Discrimination, Recruitment, Training, Tiebreak, and Preferential Treat-
ment), to which they were randomly assigned. These plans are presented in
Table 1. Respondents then rated the plan in terms of favorability to self,
favorability to their ethnic group, procedural justice, distributive justice,
and interactional justice. They also rated their intent to apply to a company
using such a plan and their overall reaction to the organization. Each of
these measures is described below. The precise items are displayed in the
appendix.
Manipulation Checks
Six items were presented to each respondent, containing one-sentence
descriptions of each of the AA plans used in this study. Note that every
respondent was instructed to rate a single plan on all six items, using a
5-point scale (1 ⫽ strongly disagree;5⫽ strongly agree). We used this
same 5-point scale for all measures described below.
Measures
Outcome unfavorability. The design of our favorability measure was
explicitly guided by two theoretical concerns. The first of these involved
the distinction between group and individual favorability. Within the
justice literature (as with much of psychology and economics) it is typical
to define outcome favorability with respect to a particular individual
respondent. This is important but limited, because individuals often judge
policies by how they impact their particular group (e.g., Mitchell, Tetlock,
Mellers, & Ordo´n˜ez, 1993; Sniderman & Tetlock, 1986). Given this, our
measure included items measuring both the individual and group favor-
ability. That having been said, we did not anticipate that these two sets of
items would be empirically distinguishable in the present research context.
While agreeing that it is proper to measure the impact of AA for both
individuals and their groups, Bobo and Kluegel (1993) observed, “The
distinction between individual and group self-interest should not be over-
drawn. Often there is a direct tie between individual self-interest and
patterns of group identification” (p. 445). Given Bobo and Kluegel’s
observations, we included items assessing both individual and group fa-
vorability but anticipated that these would load on a single factor.
A second theoretical concern had to do with the psychological impact of
potential losses. Generally speaking, individuals react more intensely to the
presence of harm than to the absence of gains (Dijksterhuis & Aarts, 2003;
Levin, Schneider, & Gaeth, 1998). Consistent with this, Brockner, Wiesen-
feld, and Martin (1995) have found that items pertaining to harm tended to
yield stronger interactions with procedural justice than did items not so
framed. Because one purpose of this present study was to assess the impact
1173
JUSTICE AND AFFIRMATIVE ACTION

of distributive justice beyond the impact of outcome favorability,we
wanted a measure of favorability that was conceptually powerful. This
renders our hypothesis tests more conservative, because it provides favor-
ability with every possible advantage in explaining variance that might
otherwise be attributable to distributive justice. For these reasons we
framed our items with an emphasis on the potentially harmful effects of
these six AA policies and thereby constructed a measure best described as
Outcome Unfavorability (e.g., “This policy takes away opportunities from
me”). All items appear in the appendix. As we will discuss in the Results
section, the scale yielded a single factor.
Organizational justice. Distributive justice was measured by two
items. An example is “This policy would give people what they deserve.”
Procedural justice was also measured with two items. An example item was
“This policy represents a procedure that is just.” Interactional justice is
generally understood to have two subparts (for reviews, see Bies, 1987,
2001; Bies & Moag, 1986; Brockner & Wiesenfeld, 1996; Sitkin & Bies,
1993). One part refers to the dignity and respect with which one is treated;
the other part refers to the availability and adequacy of the information that
is provided to the individual. The interactional justice scale measured both
of these subcomponents. Two items captured the dignity portion (e.g.,
“This policy respects the dignity of all involved”), and three items captured
the informational portion (e.g., “This policy provides all parties with the
information they need”).
Organizational attractiveness and intention to apply. Participants re-
sponded to four items written to assess the overall attractiveness of the
organization. An example item was, “I would think very highly of an
organization that selected candidates using this policy.” Intention to apply
was measured with three items taken from Highhouse, Lievens, and Sinar
(2003). An example item was “I would be interested in working for this
organization.”
Results
Manipulation Checks
With the exception of the item corresponding to the Tiebreak
Plan, for each manipulation check, participants in the appropriate
condition endorsed that item more strongly than did participants in
other conditions. Individuals in the control condition, which did
not specify AA or the elimination of discrimination, agreed that the
company’s policy involved AA (M ⫽ 2.12, SD ⫽ 1.19) to a lesser
extent than the five conditions in which an AAP was specified
(M ⫽ 3.60, SD ⫽ 1.13), t(297) ⫽⫺7.06, p ⬍ .001, d ⫽ 1.20, 95%
confidence interval (CI) for differences between conditions Ms ⫽
⫺1.92 to ⫺1.04. Individuals in the control condition (M ⫽ 3.03,
SD ⫽ 1.36) also agreed that the company’s policy involved the
elimination of discrimination to a lesser extent than the five
conditions in which elimination of discrimination was specified
(M ⫽ 3.80, SD ⫽ 1.16), t(295) ⫽⫺3.50, p ⬍ .001, d ⫽ 0.64, 95%
CI for differences between conditions Ms ⫽⫺1.28 to ⫺0.27.
Individuals in the recruitment condition (M ⫽ 3.67, SD ⫽ 0.96)
agreed that the company’s AAP involved recruitment to a greater
extent than the other five conditions (M ⫽ 3.01, SD ⫽ 1.32),
t(294) ⫽ 4.06, p ⬍ .001, d ⫽ 0.51, 95% CI for differences between
conditions Ms ⫽ 0.39 to 0.94. Individuals in the training condition
agreed that the company’s AAP used a special training program
specifically for Black applicants (M ⫽ 4.09, SD ⫽ 0.92) more than
did participants in other conditions (M ⫽ 2.36, SD ⫽ 1.09),
t(294) ⫽ 11.66, p ⬍ .001, d ⫽ 1.36, 95% CI for differences
between conditions Ms ⫽ 1.47 to 2.00. Finally, individuals in the
preferential treatment condition (M ⫽ 3.36, SD ⫽ 1.25) agreed that
the company’s AAP involved strong preferential treatment to a
greater extent than the other four conditions in which no prefer-
ential treatment was specified (M ⫽ 2.78, SD ⫽ 1.16), t(256) ⫽
2.84, p ⬍ .01, d ⫽ 0.47, 95% CI for differences between condi-
tions Ms ⫽ 0.18 to 0.98.
1
As we noted above, the only condition for which this check did
not hold was the tiebreak condition. Individuals in that condition
did not perceive that the plan involved “weak preferential treat-
ment for Black applicants” to a greater degree than the other
conditions combined, t(294) ⫽ 0.49, p ⫽ .62, d ⫽ 0.08. Because
of this result, and because the Tiebreak Plan involved preferential
treatment, we then assessed whether the tiebreak condition was
perceived to involve strong preferential treatment (the last manip-
ulation check item) to a greater extent than the control, eliminate
discrimination, recruitment, and training conditions. We found this
to be the case, as participants in the tiebreak condition (M ⫽ 3.53,
SD ⫽ 1.33) agreed with this statement to a greater extent than the
four other conditions combined (M ⫽ 2.78, SD ⫽ 1.15), t(255) ⫽
3.58, p ⬍ .001, d ⫽ 0.61, 95% CI for differences between condi-
tions Ms ⫽ 0.34 to 1.16. Therefore, participants distinguished the
tiebreak and preferential treatment conditions from all other con-
ditions but did not distinguish between the tiebreak and preferen-
tial treatment conditions, and we combined them from this point
forward in the analyses.
2
Confirmatory Factor Analysis (CFA) of Survey Measures
Though the items used in this study were based on previous
research and theory, many of them were used for the first time in
1
We also conducted analyses to ensure that there were no unexpected
differences on the manipulation check items (e.g., that individuals in the
preferential treatment condition did not perceive that there was a special
training program more than individuals in other conditions). We found only
one difference of this nature. Individuals in the preferential treatment,
tiebreak, and training conditions perceived that the AAP involved “the use
of targeted recruitment of Black applicants” more so than did those
assigned to the control condition. This is probably not surprising given the
race-conscious nature of those plans and their special emphasis on ensuring
that Black individuals are hired.
2
We had originally conceived of the Tiebreak Plan as a “Weak Prefer
-
ential Treatment Plan.” This is consistent with terminology used by
Slaughter et al. (2002). Thus, we used a manipulation check item that
measured perceptions of whether weak preferential treatment was involved
in the plan: “This company’s policy involves the use of weak preferential
treatment for Black applicants.” However, because the plan never actually
mentioned weak preferential treatment per se, it is not surprising that
participants who read this plan did not agree with this statement more than
participants in other conditions. Thus, we changed the plan label to
Tiebreak (consistent with Kravitz & Klineberg, 2000). In addition, we
conducted supplemental analyses that suggest that individuals did not
perceive the Tiebreak and Preferential Treatment Plans to differ on per-
ceptions of either “weak preferential treatment” or “strong preferential
treatment.” On the first item, “This company’s policy involves the use of
weak preferential treatment for Black applicants,” these two AAPs were
not perceived to differ significantly, t(75) ⫽⫺0.35, p ⫽ .73. On the second
item, “This company’s policy involves the use of strong preferential
treatment for Black applicants,” these two AAPs were again not perceived
to differ significantly, t(75) ⫽⫺0.57, p ⫽ .57. Thus, both plans were
viewed as involving preferential treatment (as indeed was the case), and as
we note in the text of this article, they were combined for all remaining
analyses.
1174
CROPANZANO, SLAUGHTER, AND BACHIOCHI

this present survey. Thus, we decided to examine the items using
CFA using AMOS (Version 4.0; Arbuckle & Wothke, 1999). Note
that our original model specified six correlated latent factors
(distributive justice, procedural justice, interactional justice, out-
come favorability, organizational attractiveness, and intentions to
apply) and 22 observed variables. The fit produced by this model,
however, was poor,
␹
2
(194, N ⫽ 281) ⫽ 768.25, p ⬍ .001,
Tucker–Lewis index (TLI) ⫽ .86, comparative fit index (CFI) ⫽
.88, root-mean-square error of approximation (RMSEA) ⫽ .10.
Closer inspection showed that two of the unfavorability items
(“This policy would hurt me” and “This policy would hurt people
in my ethnic group”) and one of the organizational attractiveness
items (“I would never trust a firm that selected people in this
fashion”) produced extremely high modification indices. As such,
they were lowering the overall fit of the model. Therefore, we ran
another analysis dropping these three items. This allowed us to
improve the fit of the overall model to a reasonable level:
␹
2
(137,
N ⫽ 281) ⫽ 356.71, p ⬍ .001, TLI ⫽ .93, CFI ⫽ .94, RMSEA ⫽
.073. Browne and Cudeck (1993) have suggested that models for
which the RMSEA values are .08 or less represent a reasonable
error of approximation. These factor loadings for this revised
model are presented in Table 2. Because these revisions were post
hoc, we ran all of the subsequent analyses with the original scale
and with the revised scale that lacked the aforementioned items.
The results were virtually identical, and there were no major
differences for hypothesis tests. Therefore, we report only the
analyses using the revised scales.
Correlations and Descriptive Statistics
Means, standard deviations, intercorrelations, and scale reliabil-
ity estimates are presented in Table 3. It is worth noting that the
coefficient alpha for the distributive justice measure was .55,
below Nunnally’s (1978) recommendation of .70. Although this is
lower than we would have liked, we retained both items for three
reasons. First, CFA results showed that the items loaded on a
single latent variable. Consequently, it seemed that despite the
presence of measurement error, the items were measuring the same
construct. Second, dropping either would have left us with a
one-item measure. Such a scale would likely have had dubious
psychometric properties. Third, the relatively poor reliability in
our distributive justice scale renders our analyses more conserva-
tive. Attenuation due to unreliability causes one to underestimate
population parameters (e.g., Busemeyer & Jones, 1983; Evans,
1985). Because the two items loaded on a single construct and the
presence of measurement error causes our tests to be conservative,
we retained both items in the scale.
Comparison of AAPs
Next, we compared the impact of the six plans on the dependent
variables using a multivariate analysis of variance (MANOVA).
Results of the MANOVA suggested that the AA plan manipulation
affected respondents’ perceptions, Wilks’s ⌳⫽.81, multivariate
F(6, 263) ⫽ 1.89, p ⬍ .01,
␩
2
⫽ .04. Therefore, we proceeded with
univariate analyses of variance (ANOVAs) to test the effects of
this manipulation on each dependent variable. We found that, as
expected, the various plans differed in terms of perceived distrib-
utive justice, F(5, 273) ⫽ 4.67, p ⬍ .01,
␩
2
⫽ .06; procedural
justice, F(4, 274) ⫽ 6.44, p ⬍ .001,
␩
2
⫽ .09; and interactional
justice, F(4, 273) ⫽ 3.93, p ⬍ .01,
␩
2
⫽ .05, providing at least
partial support for Hypotheses 1a, 1b, and 1c, respectively. The
manipulation also affected outcome unfavorability, F(4, 289) ⫽
2.53, p ⫽ .04,
␩
2
⫽ .03, and overall organizational attractiveness,
F(4, 272) ⫽ 3.57, p ⬍ .01,
␩
2
⫽ .06. There were no differences
between conditions on intent to apply, F(4, 272) ⫽ 1.21, p ⫽ .31,
␩
2
⫽ .02.
Table 2
Standardized Factor Loadings (
␭
x) for Full Measurement Model
Item Harm DJ PJ IJ Attraction Intentions
Takes opportunities from me 1.00
Takes opportunities from group .82
Eliminates chances for group .89
Eliminated chance to succeed .79
Give people what they deserve 1.00
Reject entitled applicants .71
Fair decision-making process 1.00
Represents just procedure .94
Valued as human beings 1.00
Respects dignity of involved .99
Easy to explain and justify .89
Provides parties with information .77
Is clear and understandable .66
Think highly of organization 1.00
Would improve my opinion .78
Socially irresponsible .44
Interested in working for organization 1.00
Would send an application .97
Probably accept a job offer .89
Note.
␹
2
(137, N ⫽ 299) ⫽ 356.71, p ⬍ .001. Tucker–Lewis index ⫽ .93, comparative fit index ⫽ .94,
root-mean-square error of approximation ⫽ .073. (The full text for each item appears in the appendix.) DJ ⫽
distributive justice; PJ ⫽ procedural justice; IJ ⫽ interactional justice.
1175
JUSTICE AND AFFIRMATIVE ACTION

Mean levels of reactions to individual plans are displayed in
Table 4. Inspection of Table 4 reveals that the Tiebreak and
Preferential Treatment Plans were rated lowest on the three justice
variables. It is interesting to note that the Eliminate Discrimination
Plan had the highest value on each of the variables. We conducted
planned contrasts, grouping the Control, Eliminate Discrimination,
and Recruitment Plans together, while also combining the Train-
ing, Tiebreak, and Preferential Treatment Plans. Race-blind plans
were rated higher on distributive justice, F ⫽ 4.62, p ⬍ .05,
␩
2
⫽
.02 (M
race-blind
⫽ 3.17, SD
race-blind
⫽ 0.85; M
race-conscious
⫽ 2.95;
SD
race-conscious
⫽ 0.82); procedural justice, F ⫽ 12.34, p ⬍ .01,
␩
2
⫽ .04 (M
race-blind
⫽ 3.27, SD
race-blind
⫽ 0.94; M
race-conscious
⫽
2.84, SD
race-conscious
⫽ 1.04); and interactional justice, F ⫽ 5.48,
p ⬍ .05,
␩
2
⫽ .02 (M
race-blind
⫽ 3.14, SD
race-blind
⫽ 0.83;
M
race-conscious
⫽ 2.90; SD
race-conscious
⫽ 0.89). Thus, Hypotheses
1a, 1b, and 1c were supported.
Relationship of Justice Perceptions to Organizational
Attractiveness and Application Intentions
To determine support for Hypotheses 2 and 3, we examined the
bivariate correlations in Table 3. Inspection of Table 3 reveals that,
as expected, distributive, procedural, and interactional justice were
each significantly related to organizational attractiveness and in-
tent to apply, suggesting support for both of our hypotheses.
However, given the high correlations among the three types of
justice, we sought to determine whether each justice perception
influenced the attraction variables, over and above each of the
other type of justice perceptions.
Therefore, we conducted six hierarchical regression analyses:
three for each dependent variable, with each type of justice entered
on the last step. In each analysis, we first controlled for outcome
unfavorability. In this way, we were able to ascertain the impact of
each type of justice beyond the impact of the other two types of
justice and outcome unfavorability. In five of the six cases, the
justice perception entered on the last step accounted for a signif-
icant proportion of the remaining variance. Procedural and inter-
actional justice each accounted for a significant amount of vari-
ance beyond the others for organizational attractiveness, F(1,
274) ⫽ 36.44, p ⬍ .001, ⌬R
2
⫽ .05, and F(1, 274) ⫽ 6.58, p ⬍ .05,
⌬R
2
⫽ .01, respectively. In predicting intentions, the incremental
effects of procedural and interactional justice were also significant,
F(1, 274) ⫽ 20.27, p ⬍ .001, ⌬R
2
⫽ .04, and F(1, 274) ⫽ 17.88,
p ⬍ .001, ⌬R
2
⫽ .03, respectively. However, distributive justice
explained incremental variance only in attractiveness, F(1, 274) ⫽
5.97, p ⬍ .05, ⌬R
2
⫽ .01, not in intentions to apply, F(1, 274) ⫽
0.42, ⌬R
2
⫽ .00.
Predicting Attractiveness Perceptions From the
Three-Way Interaction
Overview. We next investigated the hypothesized three-way
interaction in the prediction of attractiveness. Analyses proceeded
in three steps; they are shown in Table 5. First, we entered the
main effects (including outcome unfavorability); following that,
we entered the five relevant two-way interactions. The two rele-
vant three-way interactions appeared in the third step. Following
Table 3
Means, Standard Deviations, and Intercorrelations of Major Study Variables
Variable MSD 123456
1. Distributive justice 3.06 0.85 .55
2. Procedural justice 3.06 1.02 .67** .89
3. Interactional justice 3.01 0.87 .61** .77** .86
4. Outcome unfavorability 2.23 0.87 ⫺.43** ⫺.40** ⫺.39** .89
5. Organizational attractiveness 2.99 0.81 .60** .73** .68** ⫺.56** .70
6. Willingness to apply for job 3.22 0.99 .49** .65** .64** ⫺.53** .76** .92
Note. All variables were measured on 1–5 scales. Internal consistency reliabilities are reported on the diagonal.
** p ⬍ .01.
Table 4
Mean Ratings by Plan
Variable
Control
Eliminate
Discrimination Recruitment Training
Tiebreak/Preferential
Treatment
M SD M SD M SD M SD M SD
Distributive justice 2.92 0.86 3.38 0.79 3.14 0.86 3.19 0.82 2.77 0.79
Procedural justice 3.05 0.87 3.38 0.96 3.28 0.96 3.16 0.99 2.59 1.01
Interactional justice 3.01 0.70 3.21 0.80 3.15 0.89 3.15 0.89 2.70 0.85
Outcome unfavorability 2.59 0.89 1.99 0.75 2.26 0.75 2.35 0.88 2.39 0.99
Attractiveness of organization 2.77 0.79 3.21 0.76 3.18 0.76 3.06 0.79 2.73 0.86
Intention to apply 3.02 1.14 3.44 0.83 3.31 0.87 3.24 1.07 3.10 1.04
Note. Entries are means on 1–5 Likert-type scales; higher values indicate more positive ratings of the dependent variable in question.
1176
CROPANZANO, SLAUGHTER, AND BACHIOCHI

the recommendations of Aguinis (1995), we centered all of our
predictors before performing the regression analyses.
Moderated regression. Results for the moderated regression
analyses are displayed in Table 5. Overall, the model fit quite well,
F(11, 264) ⫽ 43.19, p ⬍ .001, R
2
⫽ .64. It is interesting to note
that none of the two-way interactions were significant, nor did they
collectively account for a significant increment in variance when
they were all entered at Step 2. The three-way interaction of the
three different types of justice was significant, b ⫽ .11, p ⬍ .01,
95% CI ⫽ 0.04 to 0.18. In contrast, the three-way interaction
involving outcome unfavorability was not significant, b ⫽ .03, p ⫽
.38, 95% CI ⫽⫺0.04 to 0.11. When both of these 2 three-way
interactions were entered into the equation together at Step 2, they
added a significant increment in variance. When the interaction
that involved the three types of justice was entered after all of the
aforementioned main effects and interactions, its contribution was
significant. More relevant to the present study, almost this entire
effect was produced by the three-way interaction involving justice,
⌬R
2
⫽ .01, p ⬍ .01.
Subgroup analysis. We next turned our attention to the form of
the obtained effect. Following Goldman (2003), we split the sam-
ple at the procedural justice median. We then ran regression
analyses for each group; however, the main effects and interactions
involving procedural justice were no longer used. Otherwise, we
used the same control variables as were used in the full sample.
The upper section of Table 6 displays the findings when proce-
dural justice is low. The two-way interaction between distributive
and interactional justice was significant, b ⫽⫺.16, p ⬍ .05, 95%
Table 5
Hierarchical Regression for Organizational Attractiveness
Variable
Step 1 Step 2 Step 3
bSE
␤
bSE
␤
bSE
␤
Distributive justice .08 .05 .08 .08 .05 .09 .00 .06 .00
Procedural justice .32 .05 .40** .31 .05 .40** .32 .05 .40**
Interactional justice .20 .06 .21* .20 .06 .22** .13 .06 .14
Outcome unfavorability ⫺.23 .04 ⫺.25** ⫺.23 .04 ⫺.25** ⫺.25 .05 ⫺.26**
R
2
⫽ .63**
Procedural ⫻ Distributive Justice .01 .07 .01 .01 .07 .01
Procedural ⫻ Interactional Justice .05 .04 .06 .00 .06 .01
Distributive ⫻ Interactional Justice ⫺.05 .07 ⫺.06 ⫺.06 .07 ⫺.07
Procedural Justice ⫻ Outcome Unfavorability ⫺.01 .06 ⫺.01 .00 .06 .01
Interactional Justice ⫻ Outcome Unfavorability .02 .07 .03 .06 .07 .06
R
2
⫽ .63**
⌬R
2
⫽ .00
Procedural ⫻ Interactional ⫻ Distributive .11 .04 .22**
Procedural ⫻ Interactional ⫻ Outcome Unfavorability .03 .04 .07
R
2
⫽ .64**
⌬R
2
⫽ .01**
Note. Distributive justice, procedural justice, interactional justice, and outcome unfavorability were centered for all analyses.
* p ⬍ .05. ** p ⬍ .01.
Table 6
Subgroup Analysis for Organizational Attractiveness: Low and High Procedural Justice
Variable
Step 1 Step 2
bSE
␤
bSE
␤
When procedural justice is low
Distributive justice .19 .07 .20** .11 .08 .11
Interactional justice .33 .07 .34** .23 .08 .24
Outcome unfavorability ⫺.23 .06 ⫺.29** ⫺.19 .07 ⫺.24*
Distributive Justice ⫻ Interactional Justice ⫺.16 .08 ⫺.19*
Interactional Justice ⫻ Outcome Unfavorability .07 .06 .10
When procedural justice is high
Distributive justice .11 .08 .12 .07 .12 .07
Interactional justice .41 .08 .41** .37 .11 .37**
Outcome unfavorability ⫺.29 .07 ⫺.30** ⫺.28 .11 ⫺.29**
Distributive Justice ⫻ Interactional Justice .06 .12 .08
Interactional Justice ⫻ Outcome Unfavorability ⫺.00 .13 ⫺.00
Note. Distributive justice, interactional justice, and outcome unfavorability were centered for all analyses.
* p ⬍ .05. ** p ⬍ .01.
1177
JUSTICE AND AFFIRMATIVE ACTION

CI ⫽⫺0.01 to ⫺0.30. This was not the case when procedural
justice was high (see the lower section of Table 6). Figure 1
presents a graphic representation of this interaction. The form of
this effect is quite similar to that obtained by Goldman. In partic-
ular, distributive justice has the strongest relationship (e.g., the
steepest slope) when procedural justice and interactional justice
are both low, supporting Hypothesis 4a.
Predicting Application Intentions From the Three Types
of Justice
Moderated regression. Analytic procedures for application in-
tentions paralleled those for organizational attractiveness. All pre-
dictor variables were centered before being entered in the equation.
Table 7 displays the results of our moderated regression analysis.
The predictors accounted for over half of the variance in inten-
tions, F(11, 267) ⫽ 31.71, p ⬍ .001, R
2
⫽ .56. Moreover, the
three-way interactions added a significant increment in R
2
to the
overall model. As expected, the three-way justice interaction ex-
plained significant incremental variance. In addition, the three-way
interaction involving distributive justice was statistically signifi-
cant, b ⫽ .10, p ⬍ .05, 95% CI ⫽ 0.01 to 0.19, and it contributed
unique variance even when considered beyond all of the other
variables, ⌬R
2
⫽ .01, p ⬍ .05. Unexpectedly, the three-way
interaction involving outcome unfavorability was also significant,
b ⫽ .12, p ⬍ .05, 95% CI ⫽ 0.02 to 0.22.
Subgroup analysis. Given these supportive findings, we then
explored the form of the interaction using the same procedures
used above. As shown in Table 8, the Distributive Justice ⫻
Interactional Justice interaction was significant when procedural
justice was low (b ⫽⫺.34, p ⬍ .05, 95% CI ⫽⫺0.653 to
⫺0.017), but not when procedural justice was high. Using the
techniques described earlier, we plotted the interaction in Figure 2.
Once again the findings are similar to those obtained by Goldman
(2003). When procedural justice and interactional justice are low,
the relationship of distributive justice to application intentions is
higher than it is when either (a) procedural justice is high and
interactional justice is high and (b) when procedural justice is high
and interactional justice is low. These two comparisons support
our predictions. Less consistent is the case where procedural
justice is low and interactional justice is high. As illustrated in the
upper panel of Figure 2, this slope is somewhat higher than
expected. Other than this comparison, the results support our
predicted three-way interaction.
Table 8 is also useful in that it allows us to explore the inter-
action between interactional justice and outcome unfavorability.
Notice that this effect did not achieve conventional levels of
significance, regardless of whether procedural justice was above
the median. Given these inconsistent findings, along with the fact
that this result was obtained post hoc, we did not explore this
interaction further.
Discussion
In this study we had four goals. First, we wanted to examine AA
perceptions among a group of Blacks. We believed it would be
useful to explore specifically how a group of presumed beneficia-
ries evaluated AAPs. Second, we wanted to build on existing
research that examines the content of different AAPs. Third, we
explored whether the three-way interaction involving distributive,
procedural, and interactional justice held when considering indi-
viduals’ responses to AAPs. Fourth, we sought to test this inter-
action beyond the impact of outcome unfavorability. In general,
we were successful in all of these objectives.
A central goal of our study was to emphasize the opinions of
Blacks. Blacks make up a significant proportion of the American
workforce, though historically they have been underrepresented in
social science research (e.g., Davidson & Friedman, 1998; Fried-
man & Davidson, 1999; James, 1997; Graham, 1992). We cannot
assume that findings obtained in one group will necessarily gen-
eralize to another. Indeed, prior research has identified mean level
differences in how these two groups perceive AA. The most
obvious dissimilarities are in their overall assessments; Black
Americans tend to be more supportive of AA than do White
Americans (Kravitz & Platania, 1993; Parker, Baltes, & Chris-
tiansen, 1997). In addition to these overall evaluations, there may
also be differences in the way these judgments are reached. For
example, tokenism is probably a less common experience for
White Americans than for those in other groups. Hence, Blacks
might have a better understanding of its pernicious effects. More
speculatively, members of different ethnicities may bring different
frames of reference to AA. It might be that many Whites, who
have not experienced historic racial injustice, tend to think of AA
Figure 1. Organizational attractiveness predicted by the three-way inter-
action of distributive, procedural, and interactional justice.
1178
CROPANZANO, SLAUGHTER, AND BACHIOCHI

in terms of individual-level fairness. Blacks, on the other hand,
may be more likely to evaluate AAPs in terms of their ability to
remedy group-level injustice. Although we are not aware of re-
search that has investigated this particular issue, it would certainly
be a worthwhile topic for future research. We also believe that it
will be important in future research to study some of these same
issues with (a) respondents of varying races and educational levels
and (b) multiple types of designated recipients (e.g., women,
Hispanics, or Asians). It would be especially useful to directly
compare the responses from members of different ethnicities.
Regarding the content of AA programs, our findings were
consistent with previous research (for a review see Bobocel et al.,
2001) suggesting that race-conscious AAPs evoke weaker justice
perceptions than do race-blind AAPs, supporting Hypotheses 1a,
1b, and 1c. This is likely because race-conscious AAPs contain a
tacit threat to Black applicants’ self-image (Turner & Pratkanis,
1994), by carrying the message that potential beneficiaries of AA
lack certain relevant qualifications. Race-blind AAPs, such as
those that encourage targeted recruitment, convey no such mean-
ing. Other things being equal, people tend to prefer less threatening
stimuli to more threatening stimuli.
It is worth noting that the majority of extant research suggests
that the content of race-conscious AAPs leads to disapproval by
both recipients and nonrecipients (Kravitz, 1995; Kravitz & Pla-
tania, 1993; Slaughter et al., 2002, in press). Therefore, a legiti-
mate question is whether such plans should be used at all, given the
generally unfavorable reactions that research has demonstrated.
We believe that the field needs more research on reactions to
Table 7
Hierarchical Regression for Intention to Apply for a Job
Variable
Step 1 Step 2 Step 3
bSE
␤
bSE
␤
bSE
␤
Distributive justice ⫺.04 .07 ⫺.03 ⫺.02 .07 ⫺.02 ⫺.08 .08 ⫺.07
Procedural justice .32 .07 .33** .32 .07 .32** .32 .07 .33**
Interactional justice .33 .08 .29** .34 .08 .30** .29 .08 .26**
Outcome unfavorability ⫺.35 .05 ⫺.30** ⫺.34 .06 ⫺.30** ⫺.42 .07 ⫺.36**
R
2
⫽ .54**
Procedural ⫻ Distributive Justice .18 .09 .17* .21 .09 .19*
Procedural ⫻ Interactional Justice .01 .06 .01 ⫺.01 .06 ⫺.01
Distributive ⫻ Interactional Justice ⫺.23 .10 ⫺.21* ⫺.17 .10 ⫺.15
Procedural Justice ⫻ Outcome Unfavorability .00 .08 .00 .03 .08 .03
Interactional Justice ⫻ Outcome Unfavorability ⫺.00 .09 ⫺.01 .09 .10 .08
R
2
⫽ .55**
⌬R
2
⫽ .01
Procedural ⫻ Interactional ⫻ Distributive .10 .05 .16*
Procedural ⫻ Interactional ⫻ Outcome Unfavorability .12 .05 .19*
R
2
⫽ .56**
⌬R
2
⫽ .01*
Note. Distributive justice, procedural justice, interactional justice, and outcome unfavorability were centered for all analyses.
* p ⬍ .05. ** p ⬍ .01.
Table 8
Subgroup Analysis for Application Intentions: Low and High Procedural Justice
Step 1 Step 2
bSE
␤
bSE
␤
When procedural justice is low
Distributive justice ⫺.06 .13 ⫺.04 ⫺.39 .20 ⫺.25
Interactional justice .50 .12 .36** .33 .16 .24*
Outcome unfavorability ⫺.38 .09 ⫺.37** ⫺.56 .14 ⫺.55**
Distributive Justice ⫻ Interactional Justice ⫺.34 .16 ⫺.33*
Interactional Justice ⫻ Outcome Unfavorability ⫺.16 .11 ⫺.21
When procedural justice is high
Distributive justice .06 .08 .05 .11 .10 .11
Interactional justice .41 .08 .35** .50 .10 .44**
Outcome unfavorability ⫺.34 .07 ⫺.33** ⫺.42 .09 ⫺.41**
Distributive Justice ⫻ Interactional Justice ⫺.06 .11 ⫺.07
Interactional Justice ⫻ Outcome Unfavorability .19 .12 .17
Note. Distributive justice, interactional justice, and outcome unfavorability were centered for all analyses.
* p ⬍ .05. ** p ⬍ .01.
1179
JUSTICE AND AFFIRMATIVE ACTION

race-conscious AAPs before a definitive statement can be made.
For example, the majority of research on race-based AA has
generally relied on reactions to reading the content of the AAP.
Much of Heilman’s research (e.g., Heilman et al., 1987, 1992)
suggests that women’s task performance and self-evaluation may
suffer if they are made aware that gender was a factor in selection.
Laboratory experiments examining effects of simulated race-based
selection on Black participants’ task performance are necessary
before we can draw firm conclusions about the potential stigma-
tizing effects of this type of AA. Other worthwhile topics for future
research in this area may be the effects of different types of
explanations for why the particular AAP is being used and poten-
tial differences in reactions to court-ordered versus voluntary
AAPs.
As hypothesized, we found that distributive, procedural, and
interactional justice perceptions were related to organizational
attractiveness and intentions to apply, supporting Hypotheses 2
and 3. However, when we followed up the zero-order correlations
with hierarchical regression analyses, examining the influence of
each type of justice perception after controlling for the other two
types, only procedural and interactional justice perceptions ex-
plained incremental variance in application interaction. This sug-
gests that when AAPs are being evaluated by potential Black
applicants, it is likely that the anticipated violation of procedural
justice rules and interactional injustice plays a greater role in
determining negative reactions to the organization. Although this
effect was not predicted, it is consistent with theoretical possibil-
ities raised by Folger and Cropanzano (1997, pp. 202–203). Of
course, given the post hoc nature of this finding, it should be
interpreted with caution.
As we also hypothesized, the three types of fairness interacted to
affect organizational attractiveness and intentions to apply, sup-
porting Hypothesis 4. Consistent with the prior work of Goldman
(2003) and Skarlicki and Folger (1997), the two-way interaction
between distributive justice and interactional justice predicted best
when procedural justice was low. To state the matter differently,
unfair allocations tended to have less profound effects when either
procedural justice or interactional justice was high. (There was,
however, one exception to this trend. When procedural fairness
was low and interactional fairness was high, distributive fairness
had a substantial relationship to application intentions.)
These findings suggest that individuals do not base their judg-
ments entirely on what they receive. They also examine the pro-
cess by which it is provided, as well as the interpersonal treatment
that they experience. In other words, job candidates seem to be
actively “making sense” out of their situations, combining cues
mathematically, and attempting to reach a reasoned judgment. One
might even suggest that the evaluations observed here were cau-
tious, in the sense that respondents were hesitant to form a negative
judgment until they had obtained information from multiple
sources (e.g., distributions, processes, interpersonal transactions).
This makes sense, if one considers that individuals are attempting
to reach a moral judgment concerning the fairness or appropriate-
ness of the treatment they received (cf., Cropanzano, Goldman, &
Folger, 2003; Cropanzano & Rupp, 2002; Folger & Cropanzano,
1997; Folger, Cropanzano, & Goldman, 2005). To some extent,
participants were giving an organization multiple chances to make
a “mistake” before firmly deciding that this might be an unattrac-
tive location to work.
Of course, on the basis of the present data, such a contention is
only speculative. However, there is another aspect of these find-
ings consistent with this notion. As discussed throughout, we also
included a measure of outcome unfavorability as a control variable
in all tests of main effects and interactions. Although outcome
unfavorability did indeed relate to some criteria, it did not explain
away the effects of injustice. Theoretically, this is important be-
cause it implies that the impact of justice is not only contingent on
whether an event has economically beneficial consequences. Eco-
nomic interests are vitally important to workplace decisions, of
course, but individuals may also be concerned with considerations
of social justice.
Practically speaking, our findings also have implications for
organizations and for society at large. In the design of AAPs, firms
might want to pay special attention to justice. The present study
might be helpful in this regard. In terms of all three types of
justice, participants reacted significantly favorably to the AAP that
specified recruitment. These findings are consistent with previous
research on the reactions of Black respondents (Slaughter et al.,
2002) and non-Black respondents (Kravitz, 1995). Thus, we sug-
gest that targeted minority recruitment may be the optimal strategy
for diversity enhancement (Kravitz & Platania, 1993), because it
engenders favorable reactions from both recipients and nonrecipi-
Figure 2. Intentions to apply predicted by the three-way interaction of
distributive, procedural, and interactional justice.
1180
CROPANZANO, SLAUGHTER, AND BACHIOCHI

ents and is an active attempt to identify and attract talented
minority applicants.
Of course, all studies have their limitations and this one is no
exception. One potential concern has to do with the use of vi-
gnettes. Because these scenarios were hypothetical, it could plau-
sibly be argued that participants had nothing at stake and, there-
fore, did not provide realistic answers. This concern has merit, and
future research is necessary to address this problem. However, it is
important to isolate the effects of AAPs to understand which plans
will lead to the most positive responses, and more important, why
people react differently to different plans.
It is also noteworthy that our results are consistent with prevail-
ing theory predicting a three-way interaction between the different
types of justice, as well as finding differential fairness for the
various AAPs. In addition, our findings also replicated the results
of research using different methodologies. In particular, the Gold-
man (2003) and Skarlicki and Folger (1997) studies investigated
employees’ responses to actual working conditions. This conver-
gence across different research paradigms gives us added confi-
dence in the present findings. This having been said, it should be
emphasized that the Goldman and Skarlicki and Folger papers did
not examine perceptions of AAPs. Replication of these findings in
real-world settings should be a top priority for future research
investigations.
Another potential concern could be found in the small ⌬R
2
produced by our three-way interactions (about 1% for each depen-
dent variable). These effects are smaller than the 2% found for
legal-claiming by Goldman (2003) and the 3% found for retalia-
tory behavior by Skarlicki and Folger (1997). However, in inter-
preting these small effects there are two points worth considering.
First, our modest effect sizes are within the .01–.03 range that is
common for studies of this type (Champoux & Peters, 1987;
Chaplin, 1991; Evans, 1985; McClelland & Judd, 1993). Second,
our study used a broader set of control variables than did previous
work. Specifically, we controlled for outcome favorability and all
possible interactions involving favorability. This makes our anal-
yses more conservative. Given these observations, we believe that
the three-way interaction we observed should not be ignored.
These findings should also be extended by a closer look at the
psychological mechanisms that mediate the relation between
AAPs and fairness perceptions. In our introduction, we noted the
potential for a strong AAP to stigmatize a beneficiary (e.g., Heil-
man et al., 1992, 1993; Heilman, McCullough, & Gilbert, 1996;
Heilman, Block, & Stathatos, 1997). Unfortunately, practical de-
mands for a short data collection instrument stopped us from
assessing this possibility directly. Given our results, we believe
that further investigation of this mechanism deserves high schol-
arly priority.
It also strikes us that AA provides an excellent context for
additional investigations into the meaning of distributive justice. In
the present study we only asked participants to evaluate the fair-
ness of their outcomes. We did not ask them to specify the rules
they used in making their judgment. As we discussed earlier,
Deutsch (1975, 1985) has indicated that individuals can allocate
benefits in reference to equality, equity, or need. A hypothetical
participant in our study could have been judging the distributive
fairness of these AAPs by evaluating them in accordance with any
of these rules, as well as some combination of more than one (cf.
Lissowski, Tyszka, & Okrasa, 1991). At the present time, it is not
clear which allocation rules are operable when individuals rate AA
policies, nor is it understood how evaluations of AAPs might
change depending on which allocation is salient. For example,
people who employ a need rule might prefer AAPs that are open
to people from all ethnicities so long as they are from an econom-
ically disadvantaged background. On the other hand, those that
emphasize an equity rule might prefer AAPs that pose less risk to
the economic “bottom line.” These ideas are speculative, but they
merit additional attention from scholars.
It would also be important to extend these present findings in
view of the recent research pertaining to the structure of organi-
zational justice perceptions. In this study we sought to test the
theoretical model presented by Goldman (2003) and Skarlicki and
Folger (1997). These authors divided justice into the three com-
ponents employed here—distributive, procedural, and interac-
tional. The three-way interaction tested here was predicated on this
earlier conceptual model and accompanying empirical test. Addi-
tionally, this three-component approach to justice is very widely
used and, at least according to some, remains the most popular
model in contemporary literature (e.g., Cohen-Charash & Spector,
2001). Another structural approach to organizational justice di-
vides interactional fairness into two component parts—informa-
tional justice (pertaining to explanations, social accounts, etc.) and
interpersonal justice (treating others with dignity and respect).
Research suggests that the resulting four-component structure is
very promising (e.g., Colquitt, 2001; Colquitt, Conlon, Wesson,
Porter, & Ng, 2001).
Assuming our findings are successfully replicated in future
research, they have clear implications. The content of AAPs im-
pacts perceptions of their attractiveness. In particular, our respon-
dents indicated that preferential policies were generally less fair
than other plans and were actually worse than taking no action; a
race-blind plan that still indicates some action on the part of the
organization was viewed most positively. This finding suggests
that organizations that use preferential treatment plans, which go
beyond the goal of eliminating discrimination, may be doing
themselves more harm than good in terms of attracting a diverse
applicant pool.
References
Adams, J. S. (1965). Inequity and social exchange. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 2, pp. 267–299). New
York: Academic Press.
Aguinis, H. (1995). Statistical power problems in moderated multiple regression in
management research. Journal of Management, 21, 1141–1158.
Arbuckle, J. L., & Wothke, W. (1999). Amos 4.0 user’s guide. Chicago:
SmallWaters.
Bauer, T. N., Maertz, C. P., Jr., Dolen, M. R., & Campion, M. A. (1998).
Longitudinal assessment of applicant reactions to employment testing
and test outcome feedback. Journal of Applied Psychology, 83, 892–903.
Bies, R. J. (1987). The predicament of injustice: The management of moral
outrage. In L. L. Cummings & B. M. Staw (Eds.), Research in organi-
zational behavior (Vol. 9, pp. 289–319). Greenwich, CT: JAI Press.
Bies, R. J. (2001). Interactional (in)justice: The sacred and the profane. In
J. Greenberg & R. Cropanzano (Eds.), Advances in organizational
justice (pp. 89 –118). Stanford, CA: Stanford University Press.
Bies, R. J., & Moag, J. S. (1986). Interactional justice: Communication
criteria for fairness. In B. Sheppard (Ed.), Research on negotiation in
organizations (Vol. 1, pp. 43–55). Greenwich, CT: JAI Press.
Bobo, L., & Kluegel, J. R. (1993). Opposition to race-targeting: Self-
1181
JUSTICE AND AFFIRMATIVE ACTION

interest, stratification ideology, or racial attitudes? American Sociolog-
ical Review, 58, 443– 464.
Bobocel, D. R., Davey, L. M., Son Hing, L. S., & Zanna, M. P. (2001). The
concern for justice and attractiveness of affirmative action: Cause or
rationalization? In R. Cropanzano (Ed.), Justice in the workplace: Vol.
2: From theory to practice (pp. 121–143). Mahwah, NJ: Erlbaum.
Bobocel, D. R., Son Hing, L. S., Davey, L. M., Stanley, D. J., & Zanna,
M. P. (1998). Justice-based opposition to social policies: Is it genuine?
Journal of Personality and Social Psychology, 75, 653– 669.
Brockner, J. (2002). Making sense of procedural fairness: How high
procedural fairness can reduce or heighten the influence of outcome
favorability. Academy of Management Review, 27, 58 –76.
Brockner, J., & Wiesenfeld, B. M. (1996). An integrative framework for
explaining attractiveness of decisions: The interactive effects of out-
comes and processes. Psychological Bulletin, 120, 189 –208.
Brockner, J., Wiesenfeld, B. M., & Martin, C. L. (1995). Decision frame,
procedural justice, and survivors’ reactions to job layoffs. Organiza-
tional Behavior and Human Decision Processes, 63, 59 – 68.
Browne, M. W., & Cudeck, R. (1993). Alternative ways of assessing model
fit. In K. A. Bollen & J. S. Long (Eds.), Testing structural equation
models (pp. 136 –162). Newbury Park, CA: Sage.
Busemeyer, J. R., & Jones, L. E. (1983). Analysis of multiplicative com-
bination rules when the causal variables are measured with error. Psy-
chological Bulletin, 93, 549 –562.
Cahn, S. M. (Ed.). (2002). The affirmative action debate. New York:
Routledge.
Champoux, J. E., & Peters, W. S. (1987). Form, effect size, and power in
moderated regression analysis. Journal of Occupational Psychology, 60, 243–
255.
Chaplin, W. F. (1991). The next generation of moderator research in
personality psychology. Journal of Personality, 59, 143–178.
Cohen-Charash, Y., & Spector, P. E. (2001). The role of justice in orga-
nizations: A meta-analysis. Organizational Behavior and Human Deci-
sion Processes, 86, 278 –321.
Colquitt, J. A. (2001). On the dimensionality of organizational justice: A
construct validation of a measure. Journal of Applied Psychology, 86,
386 –400.
Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C. O. L. H., & Ng,
K. Y. (2001). Justice at the millennium: A meta-analytic review of 25
years of organizational justice research. Journal of Applied Psychology,
86, 425– 445.
Cropanzano, R., Byrne, Z. S., Bobocel, D. R., & Rupp, D. R. (2001). Moral
virtues, fairness heuristics, social entities, and other denizens of organi-
zational justice. Journal of Vocational Behavior, 58, 164 –209.
Cropanzano, R., & Folger, R. (1991). Procedural justice and worker
motivation. In R. M. Steers & L. W. Porter (Eds.), Motivation and work
behavior (5th ed., pp. 131–143). New York: McGraw-Hill.
Cropanzano, R., Goldman, B. M., & Folger, R. (2003). Deontic justice:
The role of moral principles in workplace fairness. Journal of Organi-
zational Behavior, 24, 1019 –1024.
Cropanzano, R., & Greenberg, J. (1997). Progress in organizational justice:
Tunneling through the maze. In C. L. Cooper & I. T. Robertson (Eds.),
International review of industrial and organizational psychology (pp.
317–372). New York: Wiley.
Cropanzano, R., & Konovsky, M. A. (1996). Resolving the justice dilemma
by improving the outcomes: The case of employee drug screening.
Journal of Business and Psychology, 11, 239 –263.
Cropanzano, R., & Rupp, D. E. (2002). Some reflections on the morality of
organizational justice. In S. Gilliland, D. Steiner, & D. Skarlicki (Eds.),
Emerging perspectives on managing organizational justice (pp. 225–
278). Greenwich, CT: Informational Age.
Cropanzano, R., & Schminke, M. (2001). Using social justice to build
effective work groups. In M. Turner (Ed.), Groups at work: Advances in
theory and research (pp. 143–171). Hillsdale, NJ: Erlbaum.
Cropanzano, R., & Wright, T. A. (2003). Procedural justice and organiza-
tional staffing: A tale of two paradigms. Human Resource Management
Review, 13, 7–39.
Cross, T. (1994). Suppose there was no affirmative action at the most
prestigious colleges and graduate schools. Journal of Blacks in Higher
Education, 3, 44 –51.
Curry, G. E., & West, C. (Eds.). (1996). The affirmative action debate.
Cambridge, MA: Perseus.
Davidson, M., & Friedman, R. (1998). When excuses don’t work: The
persistent injustice effect among black managers. Administrative Science
Quarterly, 43, 154 –183.
Deblieux, M. (1996). Legal issues for managers: Essential skills for
avoiding your day in court. West Des Moines, IA: American Media.
Deutsch, M. (1975). Equity, equality, and need: What determines which
value will be used as the basis of distributive justice? Journal of Social
Issues, 31, 137–150.
Deutsch, M. (1985). Distributive justice: A social-psychological perspec-
tive. New Haven, CT: Yale University Press.
Dijksterhuis, A., & Aarts, H. (2003). On wildebeest and humans: The
preferential detection of negative stimuli. Psychological Sciences, 14,
14 –18.
Evans, M. G. (1985). A Monte Carlo study of the effects of correlational
method variance in moderated multiple regression. Organizational Be-
havior and Human Decision Processes, 36, 305–323.
Folger, R., & Cropanzano, R. (1997). Organizational justice and human
resources management. Thousand Oaks, CA: Sage.
Folger, R., Cropanzano, R., & Goldman, B. (2005). What is the relation-
ship between justice and morality? In J. Greenberg & J. Colquitt (Eds.),
Handbook of organizational justice (pp. 215–245). Mahwah, NJ:
Erlbaum.
Folger, R., Konovsky, M. A., & Cropanzano, R. (1992). A due process
metaphor for performance appraisal. In B. M. Staw & L. L. Cummings
(Eds.), Research in personnel and human resources management (Vol.
14, pp. 129 –177). Greenwich, CT: JAI Press.
Friedman, R., & Davidson, M. N. (1999). The Black-White gap percep-
tions of discrimination: Its causes and consequences. In R. J. Bies, R. J.
Lewicki, & B. H. Sheppard (Eds.), Research on negotiation in organi-
zations (Vol. 7, pp. 203–228). Greenwich, CT: JAI Press.
Gilliland, S. W. (1993). The perceived fairness of selection systems: An
organizational justice perspective. Academy of Management Review, 18,
694 –734.
Gilliland, S. W., & Steiner, D. D. (2001). Causes and consequences of
applicant perceptions of unfairness. In R. Cropanzano (Ed.), Justice in
the workplace (Vol. 2): From theory to practice (pp. 175–195). Mahwah,
NJ: Erlbaum.
Goldman, B. M. (2003). The application of referent cognitions theory to
legal-claiming by terminated workers: The role of organizational justice
and anger. Journal of Management, 29, 705–728.
Graham, S. (1992). Most of the subjects were White and middle class:
Trends in published research on African Americans in selected APA
journals, 1970 –1989. American Psychologist, 47, 629 – 639.
Heilman, M., & Alcott, V. B. (2001). What I think you think of me:
Women’s reactions to being viewed as beneficiaries of preferential
selection. Journal of Applied Psychology, 86, 574 –582.
Heilman, M., Block, C. J., & Stathatos, P. (1997). The affirmative action
stigma of incompetence: Effects of performance information ambiguity.
Academy of Management Journal, 40, 603–625.
Heilman, M. E., Block, C. J., & Lucas, J. A. (1992). Presumed incompe-
tent? Stigmatization and affirmative action efforts. Journal of Applied
Psychology, 77, 536 –544.
Heilman, M. E., Kaplow, S. R., Amato, M. A., & Stathatos, P. (1993).
When similarity is a liability: Effects of sex-based preferential selection
on reactions to like-sex and different-sex others. Journal of Applied
Psychology, 78, 917–927.
1182
CROPANZANO, SLAUGHTER, AND BACHIOCHI

Heilman, M. E., McCullough, W. F., & Gilbert, D. (1996). The other side
of affirmative action: Reactions of nonbeneficiaries to sex-based pref-
erential selection. Journal of Applied Psychology, 81, 346 –357.
Heilman, M. E., Simon, M. C., & Repper, D. P. (1987). Intentionally
favored, unintentionally harmed? Impact of sex-based preferential se-
lection on self-perceptions and self-evaluations. Journal of Applied
Psychology, 72, 62– 68.
Highhouse, S., Lievens, F., & Sinar, E. F. (2003). Measuring attraction to
organizations. Educational and Psychological Measurement, 63, 986 –1001.
James, K. (1997). Social identity and health-related costs for organizations:
A comparative study of minority and majority workers. Journal of
Occupational Health Psychology, 2, 108 –117.
Judge, T. A., & Bretz, R. D. (1992). Effects of work values on job choice
decisions. Journal of Applied Psychology, 77, 261–271.
Kravitz, D. A. (1995). Attitudes toward affirmative action plans directed at
Blacks: Effects of plan and individual differences. Journal of Applied
Social Psychology, 25, 2192–2220.
Kravitz, D. A., Harrison, D. A., Turner, M. E., Levine, E. L., Chaves, W.,
Brannick, M. T., et al. (1997). Affirmative action: A review of psycho-
logical and behavioral research. Bowling Green, OH: Society for In-
dustrial and Organizational Psychology.
Kravitz, D. A., & Klineberg, S. L. (2000). Attractiveness of two versions
of affirmative action among Whites, Blacks, and Hispanics. Journal of
Applied Psychology, 85, 597– 611.
Kravitz, D. A., & Platania, J. (1993). Attitudes and beliefs about affirma-
tive action: Effects of target and of respondent sex and ethnicity. Journal
of Applied Psychology, 78, 928 –938.
Leck, J. D., Saunders, D. M., & Charbonneau, M. (1996). Affirmative
action programs: An organizational justice perspective. Journal of Or-
ganizational Behavior, 17, 79 – 89.
Leventhal, G. S. (1976). Fairness in social relationships. In J. W. Thibaut,
J. T. Spence, & R. C. Carson (Eds.), Contemporary topics in social
psychology (pp. 211–240). Morristown, NJ: General Learning Press.
Levin, I. P., Schneider, S. L., & Gaeth, G. J. (1998). All frames are not
created equal: A typology and critical analysis of framing effects.
Organizational Behavior and Human Decision Processes, 76, 149 –188.
Lind, E. A. (2001). Fairness heuristic theory: Justice judgments as pivotal
cognitions in organizations. In J. Greenberg & R. Cropanzano (Eds.),
Advances in organizational justice (pp. 56 – 88). Stanford, CA: Stanford
University Press.
Lissowski, G., Tyszka, T., & Okrasa, W. (1991). Principles of distributive
justice: Experiments in Poland and America. Journal of Conflict Reso-
lution, 35, 98 –119.
McClelland, G. H., & Judd, C. M. (1993). Statistical difficulties of detect-
ing interactions and moderator effects. Psychological Bulletin, 114,
376 –390.
Messick, D. M., & Sentis, K. (1979). Fairness and preference. Journal of
Experimental Social Psychology, 15, 418 –434.
Messick, D. M., & Sentis, K. (1983). Fairness, preference, and fairness
biases. In D. M. Messick & K. S. Cook (Eds.), Equity theory: Psycho-
logical and sociological perspectives (pp. 61–94). New York: Praeger.
Mitchell, G., Tetlock, P. E., Mellers, B. A., & Ordo´n˜ez, L. D. (1993).
Judgments of social justice: Compromises between equality and effi-
ciency. Journal of Personality and Social Psychology, 65, 629 – 639.
Mosley, A. G., & Capaldi, N. (1996). Affirmative action: Social justice or
unfair preference? Lanham, MD: Rowman & Littlefield.
Nacoste, R. W. (1987). But do they care about fairness? The dynamics of
preferential treatment and minority interest. Basic and Applied Social
Psychology, 8, 177–191.
National Science Foundation, Division of Science Resources Statistics.
(2003). Women, minorities, and persons with disabilities in science and
engineering: 2002 (Rep. No. NSF 03–312). Arlington, VA: Author.
Nunnally, J. C. (1978). Psychometric theory (2nd ed.). New York:
McGraw-Hill.
Parker, C. P., Baltes, B. B., & Christiansen, N. D. (1997). Support for
affirmative action, justice perceptions, and work attitudes: A study of
gender and racial-ethnic group differences. Journal of Applied Psychol-
ogy, 82, 376 –389.
Robinson, R. K., Seydel, J., & Douglas, C. (1998). Affirmative action: The
facts, the myths, and the future. Employee Responsibilities and Rights
Journal, 11, 99 –115.
Singer, M. S. (1992). Procedural justice in managerial selection: Identifi-
cation of fairness determinants and consequences of fairness percep-
tions. Social Justice Research, 5, 49–70.
Singer, M. S. (1993). Fairness in personnel selection: An organizational
justice perspective. Aldershot, New Zealand: Avebury.
Singer, M. S. (1996). Effects of scope of justice, informant ethnicity, and
information frame on attitudes towards ethnicity-based selection. Inter-
national Journal of Psychology, 31, 191–205.
Sitkin, S. B., & Bies, R. J. (1993). Social accounts in conflict settings.
Human Relations, 46, 349 –370.
Skarlicki, D. P., & Folger, R. (1997). Retaliation in the workplace: The
roles of distributive, procedural, and interactional justice. Journal of
Applied Psychology, 82, 434 – 443.
Skitka, L. J., Winquist, J., & Hutchinson, S. (2003). Are outcome fairness
and outcome favorability distinguishable psychological constructs? A
meta-analytic review. Social Justice Research, 16, 309 –341.
Skrentny, J. D. (1996). The ironies of affirmative action: Politics, culture,
and justice in America. Chicago: University of Chicago Press.
Slaughter, J. E., Bulger, C. A., & Bachiochi, P. D. (in press). Black
applicants’ reactions to affirmative action plans: Effects of perceived
procedural fairness, anticipated stigmatization, and anticipated remedia-
tion of injustice. Journal of Applied Social Psychology.
Slaughter, J. E., Sinar, E. F., & Bachiochi, P. D. (2002). Black applicants’ reactions
to affirmative action plans: Effects of plan content and previous experience with
discrimination. Journal of Applied Psychology, 87, 333–344.
Sniderman, P., & Tetlock, P. E. (1986). Public opinion and political
ideology. In M. Herman (Eds.), Political psychology (pp. 62–96). San
Francisco: Jossey-Bass.
Sovereign, K. L. (1994). Personnel law (3rd ed.). Englewood Cliffs, NJ:
Prentice Hall.
Sowell, T. (1990). Preferential policies: An international perspective. New
York: William Morrow and Company.
Thibaut, J. W., & Kelley, H. H. (1959). The social psychology of groups.
New York: Wiley.
Turner, M. E., & Pratkanis, A. R. (1994). Affirmative action as help: A
review of recipient attractiveness of preferential selection and affirma-
tive action. Basic and Applied Social Psychology, 15, 43– 69.
Tyler, T. R., & Blader, S. L. (2000). Cooperation in groups: Procedural
justice, social identity, and behavioral engagement. Philadelphia: Psy-
chology Press.
Tyler, T. R., & Caine, A. (1981). The influence of outcomes and proce-
dures on satisfaction with formal leaders. Journal of Personality and
Social Psychology, 41, 642– 655.
Van den Bos, K. (1998, May). Trying to understand the psychology of
procedural and distributive justice by means of fairness heuristic theory.
Paper presented at the Seventh International Conference on Social
Justice Research, Denver, Colorado.
Van den Bos, K., Wilke, H. A. M., Lind, E. A., & Vermunt, R. (1998).
Evaluating outcomes by means of the fair process effect: Evidence for
different processes in fairness and satisfaction judgments. Journal of
Personality and Social Psychology, 74, 1493–1503.
Velasquez, M. G. (1992). Business ethics: Concepts and cases (3rd ed.).
Englewood Cliffs, NJ: Prentice Hall.
Weiss, H. M., Suckow, K., & Cropanzano, R. (1999). Effects of justice
conditions on discrete emotions. Journal of Applied Psychology, 84,
786 –794.
1183
JUSTICE AND AFFIRMATIVE ACTION
(Appendix follows)

Appendix
Survey Instruments Used in This Study
These items were used to measure the constructs examined in this study. All of these scales were completed using the following 5-point scale: 1 ⫽
strongly disagree,2⫽ slightly disagree,3⫽ neutral,4⫽ slightly agree,5⫽ strongly agree.
Manipulation Checks
This company’s policy involves:
. . . an affirmative action plan.
. . . the elimination of discrimination against Black applicants.
. . . the use of targeted recruitment of Black applicants.
. . . the use of a special training program for Black applicants.
. . . the use of weak preferential treatment for Black applicants.
. . . the use of strong preferential treatment for Black applicants.
Distributive Justice
1. This policy would reject many applicants who are entitled to jobs
[reverse coded].
2. This policy would give people what they deserve.
Procedural Justice
1. This policy represents a procedure that is just.
2. This policy establishes a fair decision-making process.
Interactional Justice
1. This policy respects the dignity of all involved.
2. This policy shows that all applicants are valued as human beings.
3. This policy provides all parties with the information they need.
4. This policy is clear and understandable to all.
5. This policy is easy to explain and justify.
Outcome Unfavorability
1. This policy virtually eliminates my chance to succeed.
2. This policy would hurt me. [This item was dropped due to poor
fit.]
3. This policy takes opportunities away from me.
4. This policy virtually eliminates the chance to succeed for many
people in my ethnic group.
5. This policy would hurt people in my ethnic group. [This item was
dropped due to poor fit.]
6. This policy takes opportunities away from people in my ethnic
group.
Application Intentions
1. I would be interested in working for this organization.
2. I would send an application to this organization.
3. I would probably accept a job offer from this organization.
Organizational Attractiveness
1. I would think very highly of an organization that selected can-
didates using this policy.
2. An organization that uses this policy is likely to be socially
irresponsible.
3. I would never trust a firm that selected people in this fashion.
[This item was dropped due to poor fit.]
4. If I learned that a firm used this policy, it would improve my
opinion of them.
Received January 14, 2004
Revision received December 16, 2004
Accepted December 20, 2004 䡲
1184
CROPANZANO, SLAUGHTER, AND BACHIOCHI