Work Group Diversity and Group Performance: An Integrative Model and Research Agenda

Abstract

Research on the relationship between work group diversity and performance has yielded inconsistent results. To address this problem, the authors propose the categorization-elaboration model (CEM), which reconceptualizes and integrates information/decision making and social categorization perspectives on work-group diversity and performance. The CEM incorporates mediator and moderator variables that typically have been ignored in diversity research and incorporates the view that information/decision making and social categorization processes interact such that intergroup biases flowing from social categorization disrupt the elaboration (in-depth processing) of task-relevant information and perspectives. In addition, the authors propose that attempts to link the positive and negative effects of diversity to specific types of diversity should be abandoned in favor of the assumption that all dimensions of diversity may have positive as well as negative effects. The ways in which these propositions may set the agenda for future research in diversity are discussed.

Full text

Work Group Diversity and Group Performance: An Integrative Model and
Research Agenda
Daan van Knippenberg
Erasmus University Rotterdam
Carsten K. W. De Dreu and Astrid C. Homan
University of Amsterdam
Research on the relationship between work group diversity and performance has yielded inconsistent
results. To address this problem, the authors propose the categorization-elaboration model (CEM), which
reconceptualizes and integrates information/decision making and social categorization perspectives on
work-group diversity and performance. The CEM incorporates mediator and moderator variables that
typically have been ignored in diversity research and incorporates the view that information/decision
making and social categorization processes interact such that intergroup biases flowing from social
categorization disrupt the elaboration (in-depth processing) of task-relevant information and perspectives.
In addition, the authors propose that attempts to link the positive and negative effects of diversity to
specific types of diversity should be abandoned in favor of the assumption that all dimensions of diversity
may have positive as well as negative effects. The ways in which these propositions may set the agenda
for future research in diversity are discussed.
Work-group diversity is a fact of organizational life. It is also a
key concern for theory and practice in organizational behavior.
Groups in organizations have become more diverse in terms of
their demographic composition over the years and will continue to
become more diverse in years to come (Jackson, 1992; Triandis,
Kurowski, & Gelfand, 1994; Williams & O’Reilly, 1998). In
addition, organizations are increasingly turning to the use of cross-
functional teams, thus increasingly introducing functional diver-
sity in work groups. Because work-group diversity may have
positive as well as negative effects on group performance (e.g.,
Guzzo & Dickson, 1996; Milliken & Martins, 1996; Williams &
O’Reilly, 1998), the questions of which processes underlie these
effects of diversity and how to manage these processes pose major
challenges to organizational theory and practice.
Unfortunately, research on the positive and negative effects of
work-group diversity has largely developed in separate research
traditions, and an integrative theoretical framework from which to
understand the effects of diversity on group performance is miss-
ing (Guzzo & Dickson, 1996; Kozlowski & Bell, 2003; Williams
& O’Reilly, 1998). Indeed, even though contemporary models of
diversity seem able to explain the positive or negative effects of
diversity when they occur, they are less able to predict when each
occurs. As a case in point, recent attempts at meta-analytical
integration have not been very successful in linking diversity with
performance (Bowers, Pharmer, & Salas, 2000; Webber &
Donahue, 2001; also see Wood, 1987). To meet this challenge and
to advance our understanding of the effects of work-group diver-
sity on group performance, we introduce a model of the processes
underlying the positive and the negative effects of diversity that we
believe has greater predictive power and opens up new directions
in research on diversity and group performance. We begin with a
brief review of the state of the art in diversity research to position
our model and the problem it aims to address. We then introduce
our model and discuss its core propositions, highlighting their
implications for diversity research.
State of the Art: A Brief Review of Work-Group
Diversity and Performance
Diversity refers to differences between individuals on any at-
tribute that may lead to the perception that another person is
different from self (e.g., Jackson, 1992; Triandis et al., 1994;
Williams & O’Reilly, 1998). In principle, diversity thus refers to
an almost infinite number of dimensions, ranging from age to
nationality, from religious background to functional background,
from task skills to relational skills, and from political preference to
sexual preference. In practice, however, diversity research has
mainly focused on gender, age, race/ethnicity, tenure, educational
background, and functional background (Milliken & Martins,
1996; Williams & O’Reilly, 1998). A number of researchers have
proposed that the most important difference underlying diversity
dimensions is that between social category diversity— differences
in readily detectable attributes such as sex, age, and ethnicity—and
informational/functional diversity— differences in less visible un-
derlying attributes that are more job-related, such as functional and
educational background (Jackson, 1992; Jehn, Northcraft, &
Neale, 1999; Milliken & Martins, 1996; Tsui, Egan, & O’Reilly,
1992).
Daan van Knippenberg, Rotterdam School of Management, Erasmus
University Rotterdam, Rotterdam, the Netherlands; Carsten K. W. De Dreu
and Astrid C. Homan, Department of Psychology, University of Amster-
dam, Amsterdam, the Netherlands.
We are indebted to Barbara van Knippenberg, Bernard Nijstad, Frederic
Damen, Laurens Rook, and Wendy van Ginkel for their valuable comments
on drafts of this article.
Correspondence concerning this article should be addressed to Daan van
Knippenberg, Rotterdam School of Management, Erasmus University Rot-
terdam, P.O. Box 1738, 3000 DR, Rotterdam, the Netherlands. E-mail:
dvanknippenberg@fbk.eur.nl
Journal of Applied Psychology Copyright 2004 by the American Psychological Association
2004, Vol. 89, No. 6, 1008 –1022 0021-9010/04/$12.00 DOI: 10.1037/0021-9010.89.6.1008
1008

In a comprehensive review of the literature, Williams and
O’Reilly (1998) identified two main traditions in research in
work-group diversity and performance: the social categorization
perspective and the information/decision-making perspective. The
social categorization perspective holds that similarities and differ-
ences are used as a basis for categorizing self and others into
groups, with ensuing categorizations distinguishing between one’s
own in-group and one or more out-groups. People tend to like and
trust in-group members more than out-group members and thus
generally tend to favor in-groups over out-groups (Brewer, 1979;
Tajfel & Turner, 1986; Turner, Hogg, Oakes, Reicher, & Weth-
erell, 1987). Consistent with research on similarity/attraction (Wil-
liams & O’Reilly, 1998), this signifies that work group members
are more positively inclined toward their group and the people
within it if fellow group members are similar rather than dissimilar
to the self. Moreover, categorization processes may produce sub-
groups within the work group (i.e., “us” and “them”), and give rise
to problematic inter-subgroup relations. As a result, the more
homogeneous the work group, the higher member commitment
(Riordan & Shore, 1997; Tsui et al., 1992) and group cohesion
(O’Reilly, Caldwell, & Barnett, 1989) will be, the fewer relational
conflicts will occur (Jehn et al., 1999; Pelled, Eisenhardt, & Xin,
1999), and the less likely membership will be to turn over (Wag-
ner, Pfeffer, & O’Reilly, 1984). Together, these processes are
proposed to result in higher overall group performance when
groups are homogeneous rather than heterogeneous (for evidence,
see Jehn et al., 1999; Murnighan & Conlon, 1991; Simons, Pelled,
& Smith, 1999).
The information/decision-making perspective arrives at quite
different predictions, holding that diverse groups should outper-
form homogeneous groups. The idea is that diverse groups are
more likely to possess a broader range of task-relevant knowledge,
skills, and abilities that are distinct and nonredundant and to have
different opinions and perspectives on the task at hand. This not
only gives diverse groups a larger pool of resources, but may also
have other beneficial effects. The need to reconcile conflicting
viewpoints may force the group to more thoroughly process task-
relevant information and may prevent the group from opting too
easily for a course of action on which there seems to be consensus.
In addition, exposure to diverging and potentially surprising per-
spectives may lead to more creative and innovative ideas and
solutions (Ancona & Caldwell, 1992; Bantel & Jackson, 1989; De
Dreu & West, 2001). Corroborating this analysis, some studies
found an association of diversity with increased task conflict (Jehn
et al., 1999; Pelled et al., 1999) and higher performance and
innovation (Bantel & Jackson, 1989; Cox, Lobel, & McLeod,
1991; Jehn et al., 1999).
Whereas the social categorization perspective focuses more on
relational aspects, the information/decision-making perspective
concentrates on task-related aspects of group processes. One could
thus argue that diversity negatively affects relationships within the
group while simultaneously contributing to group performance
(Triandis et al., 1994). However, because problematic intragroup
relations and low group cohesiveness are detrimental to perfor-
mance (De Dreu & Weingart, 2003b; Jehn, 1995; Mullen &
Copper, 1994), it is difficult to see how diversity could negatively
affect relationships while at the same time stimulate performance.
An alternative way to integrate social categorization and
information/decision-making perspectives is to take the dimen-
sions of diversity into account. Social categorization theories out-
line how categorizations are more likely to be used if category
markers are readily detectable, such as in the case of differences in
sex and race (Fiske, 1998). Moreover, categorizations that are
more cognitively accessible are more likely to be used. This
renders categorizations that people have been socialized into using
throughout their life (i.e., such as sex, age, and race/ethnicity)
particularly likely to be used (A. van Knippenberg & Dijksterhuis,
2000). Accordingly, one may argue that the social categorization
effects of diversity should be more likely to occur for such readily
observable characteristics as sex and ethnicity (i.e., social category
diversity) than for more “hidden” dimensions of diversity such as
functional or educational background. The positive effects of di-
versity, in contrast, are linked to differences in information, ex-
pertise, viewpoints, and so forth. Thus, one may argue that diver-
sity only has positive effects on group performance to the extent
that it covaries with informational differences. Indeed, it has been
proposed that although demographic differences in, for instance,
sex, age, and ethnicity may be associated with such informational
differences (Cox et al., 1991; Tsui & O’Reilly, 1989), more
job-related dimensions such as functional and educational back-
ground (i.e., informational diversity) are more likely to be associ-
ated with (job-relevant) informational differences (Pelled et al.,
1999). The positive effects of diversity, as outlined in the
information/decision-making perspective, therefore, would be
more likely to occur for diversity on underlying job-related
attributes.
Although the logic behind this reasoning seems sound, there is
insufficient empirical support for the proposition that the effects of
diversity are contingent on diversity type (i.e., social category vs.
informational diversity). Some studies have yielded evidence con-
sistent with this proposition (e.g., Jehn et al., 1999), whereas others
revealed positive performance effects of demographic differences
(e.g., Cox et al., 1991) or negative performance effects of differ-
ences in job-related attributes (e.g., Simons et al., 1999). More-
over, recent meta-analyses by Bowers et al. (2000) and Webber
and Donahue (2001) failed to support the proposition that diversity
type moderates the effects of diversity on performance. These
studies showed that neither diversity on readily observable at-
tributes nor diversity on underlying job-related attributes could be
reliably linked to group performance.
The Categorization–Elaboration Model
Previous research and theory thus has not been able to ade-
quately account for the positive and negative effects diversity in
work groups can have or to integrate the social categorization and
information/decision-making perspective in a satisfactory way. To
address this problem, we propose the categorization– elaboration
model (CEM). The CEM identifies a number of reasons why
diversity research has yielded such inconsistent findings, each of
which is discussed in more detail in the following sections. First,
diversity research has paid insufficient attention to group informa-
tion processing, and to important moderators of group information
processing, as the process underlying the positive effects of diver-
sity. Second, diversity research has often worked from a somewhat
oversimplified conceptualization of social categorization pro-
cesses. This has apparently led diversity research to largely ignore
important moderators of the relationships between diversity and
social categorization and between social categorization and the
negative consequences of categorization (i.e., intergroup bias).
1009
SPECIAL SECTION: WORK GROUP DIVERSITY

Third, diversity research has typically studied information/
decision-making processes and social categorization processes in
isolation, whereas the CEM suggests that information/decision
making and social categorization processes interact. Finally, much
diversity research has worked from the assumption that
information/decision-making and social categorization processes
are each associated with particular dimensions of diversity. Our
reconsideration of the nature of these processes suggests, however,
that each dimension of diversity may in principle elicit both
information/decision-making and social categorization processes.
This reconceptualization and integration of information/decision-
making and social categorization perspectives in the CEM points
to mediator and moderator variables that typically have been
ignored in diversity research and suggests a research agenda for
future research in diversity and group performance that may re-
solve many of the inconsistent findings in diversity research.
The CEM is depicted in Figure 1. Building on the information/
decision-making perspective, we propose that diversity within a
group is positively related to the elaboration of task-relevant
information and perspectives within the group—that is, to group
members’ exchange, discussion, and integration of ideas, knowl-
edge, and insights relevant to the group’s task. Elaboration of
task-relevant information and perspectives, in turn, is proposed to
be related to group performance, especially to group creativity,
innovation, and decision quality. This is not to say, however, that
diversity within a group will always lead to elaboration of task-
relevant information and perspectives. Accordingly, we identify
factors that moderate the relationship between diversity and elab-
oration. In brief, we propose that diversity in a group is most likely
to lead to elaboration of task-relevant information and perspectives
when the group task has strong information-processing and
decision-making components, when the group is highly motivated
to process task-relevant information and perspectives, and when
group members are high in task ability.
Consistent with the social categorization perspective in diversity
research, we also attribute a role to social categorization (the
differentiation between in-group others, who are subjectively sim-
ilar to self, and outgroup others, who are subjectively dissimilar to
self). Building on theory and research in social categorization, we
propose that the extent to which differences between group mem-
bers engender social categorization is contingent on three factors:
the cognitive accessibility, the normative fit, and the comparative
fit of the categorization. Cognitive accessibility refers to the ease
with which the social categorization implied by the differences
(e.g., men vs. women) is cognitively activated. Normative fit
reflects the extent to which the categorization makes subjective
sense to group members. Comparative fit refers to the extent to
which the categorization yields subgroups with high intragroup
similarity and high intergroup differences. We also note that social
categorization per se should not be equated with intergroup biases
(more favorable responses to others categorized as in-group than
others categorized as out-group) that may flow from social cate-
gorization and argue that intergroup bias and not social categori-
zation per se is disruptive to diverse groups. A key question thus
Figure 1. The categorization-elaboration model of work group diversity and group performance. Superscripts
refer to the correspondingly numbered propositions discussed in the text.
1010
VAN KNIPPENBERG, DE DREU, AND HOMAN

is under what conditions social categorization results in intergroup
bias. We propose that threats and challenges to subgroup identity
are the main factors driving intergroup bias.
Integrating the information/decision-making and social catego-
rization perspectives that have developed as separate traditions, we
propose that elaboration and social categorization processes inter-
act in that the intergroup biases that may result from social cate-
gorization disrupt elaboration of task-relevant information and
perspectives. Finally, in contrast to earlier attempts to link the
positive and negative effects of diversity to specific types of
diversity, we propose that all dimensions of diversity may elicit
elaboration of task-relevant information as well as social catego-
rization processes.
Each of these propositions is discussed in the following sections.
This discussion highlights how these propositions help resolve
inconsistencies among findings in diversity research and may
assist in setting the agenda for future research. First, we discuss
our propositions regarding diversity and the elaboration of task-
relevant information. Next, we take a closer look at the relation-
ship between diversity, social categorization, and intergroup bias.
Integrating these first two sections, we follow with a discussion of
the ways in which intergroup biases may disrupt the elaboration of
task-relevant information and perspectives in diverse groups. Fi-
nally, we outline the ways in which all dimensions of diversity
may engender elaboration as well as social categorization pro-
cesses. To further clarify our propositions and highlight the re-
search implications, we conclude each section with a discussion of
implications for research. We end with the applications our model
might have for diversity management.
Diversity and the Elaboration of Task-Relevant
Information
Part of the potential advantage of diverse groups over homoge-
neous groups lies in the greater pool of task-relevant information
and expertise diverse groups may have at their disposal. In addi-
tion, the larger social network of diverse groups may give diverse
groups more access to new information and a potentially larger
basis of support for decisions than homogeneous groups (Ancona
& Caldwell, 1992). Our concern here, however, is with the group
processes engendered by diversity. Arguably, these processes are
the key to realizing diversity’s potential, because it is not the
availability of information per se but the use of this information in
group task performance that lies at the basis of diverse groups’
potentially superior performance. With regard to these processes,
however, we identify two problems in the diversity literature. First,
the processes underlying the positive effects of diversity are typ-
ically conceptualized and operationalized in terms of conflict and
dissent, but the available evidence does not support a positive
relationship between conflict and performance (De Dreu & Wein-
gart, 2003b). Second, and closely related to the first concern, the
nature of the performance that is expected to benefit from diversity
is ill-defined.
The information/decision-making processes that are proposed to
underlie the positive effects of diversity are typically operational-
ized as conflict or dissent (e.g., Jehn et al., 1999; Pelled et al.,
1999). Differences in information and viewpoints may give rise to
task conflict and dissent; faced with the need to solve these
conflicts and reconcile opposing views, group members may en-
gage in more elaborate processing of task-relevant information and
search for more creative problem solutions than would be the case
in the absence of conflict and dissent. As a result, task conflict and
dissent may be associated with better and more creative group
performance (e.g., De Dreu, Harinck, & van Vianen, 1999; Tjos-
vold, 1998). Although this analysis holds intuitive appeal, a num-
ber of considerations suggest that conflict and dissent are at best
only a proxy of, or a precursor to, the processes that underlie the
positive effect of diversity on group performance.
First, a meta-analysis of the conflict literature shows that task
conflict is negatively related to group performance (De Dreu &
Weingart, 2003b). Second, it is not so much the presence or
absence of conflict but instead the way conflicts are managed that
helps or hinders teams to perform effectively (De Dreu & Wein-
gart, 2003a; Lovelace, Shapiro, & Weingart, 2001; Simons &
Peterson, 2000; Tjosvold, 1998). For instance, when group mem-
bers decide to avoid rather than confront conflict of opinions, the
potentially beneficial effects of diversity will not be harvested.
Third, and perhaps most important, performance does not benefit
from conflict and dissent per se but from the process that conflict
and dissent is assumed to promote: the deep-level and creative
processing of diverse information and viewpoints. However, the
exchange of diverse information, ideas, and viewpoints may also
stimulate such in-depth processing without conflict or dissent (i.e.,
one may process new information without initially disagreeing
with the position implied by the information). It is possible that
under certain conditions conflict and dissent may promote this
process, but as suggested above, it may also disrupt this process
(e.g., when groups opt for a quick compromise to avoid the
conflict). The bottom line therefore is that conflict and dissent are
not necessary to realize the potential benefits of diversity nor are
they always conducive to realizing these benefits.
Accordingly, we propose that elaboration of task-relevant infor-
mation is the primary process underlying the positive effects of
diversity on performance. Building on the conceptualization of
groups as information processors (Hinsz, Tindale, & Vollrath,
1997), elaboration is defined as the exchange of information and
perspectives, individual-level processing of the information and
perspectives, the process of feeding back the results of this
individual-level processing into the group, and discussion and
integration of its implications. It is this elaboration, then, that may
be engendered by work-group diversity, and that may lead diverse
groups to outperform more homogeneous groups. Consider, for
example, an information technology company that relies on cross-
functional project teams consisting of designers, programmers, and
sales and after-sales representatives to develop and sell tailor-made
software packages. High-quality performance requires that team
members inform the team on the basis of their own expertise about
the different issues involved (e.g., customer wishes, design and
program possibilities, costs), carefully process the perspectives
introduced by other team members to understand the implications
for their own area of expertise, feed these implications back to the
team, and through integration of perspectives design the optimal
product.
First evidence for the role of elaboration in diverse groups
comes from a recent experimental study of decision making in
informationally diverse groups by van Ginkel and van Knippen-
berg (2003). In this study, the extent to which group members had
a shared representation of the task (i.e., a shared mental model;
Cannon-Bowers & Salas, 2001) that stressed the sharing and
processing of diverse information and viewpoints (i.e., elaboration
1011
SPECIAL SECTION: WORK GROUP DIVERSITY

of task-relevant information) was manipulated through task in-
structions and group discussion of these instructions. Results
showed that groups with shared mental models emphasizing elab-
oration made higher quality decisions than groups that held such
shared task representations to a lesser extent.
Proposition 1: The primary process underlying the positive
effects of diversity on group performance is elaboration of
task-relevant information.
Moderators of the Relationship Between Diversity and
Elaboration
A focus on elaboration of task-relevant information as the
primary process driving the positive effects of diversity points to a
number of important moderator variables that probably flow less
obviously from an emphasis on conflict and dissent or from other,
more loosely defined approaches to information/decision-making
processes. As discussed in the previous section, the proposition
that diverse groups may outperform more homogeneous groups
follows from the reasoning that the exposure to more diverse
information and perspectives may promote elaboration of task-
relevant information. This, in turn, would be expected to lead to
more thorough and creative information processing, problem solv-
ing, and decision making. Such expected outcomes, then, give rise
to the proposition that diversity may benefit performance to the
extent that performance requires information processing, creative
and innovative idea generation and problem solving, and/or high-
quality decision making. That is, group performance should ben-
efit primarily from diversity on complex and nonroutine
information-processing and decision-making tasks where perfor-
mance is primarily defined in terms of the quality of the group’s
product, such as those facing research and development teams.
Indeed, such tasks may be expected to invite more elaborate
information processing in the first place (Stewart & Barrick, 2000;
Van de Ven, Delbecq, & Koenig, 1976), setting the stage for the
potentially positive effects of diversity of information and perspec-
tives. In contrast, there appears little reason to expect that simple
and routine tasks (e.g., repetitive production tasks) typically invite
extensive information processing. In fact, elaboration of task in-
formation in routine tasks may result in abandoning otherwise
reasonable work procedures and thus may be counterproductive
(De Dreu & Weingart, 2003b; Schwenk, 1990). Task information-
processing and decision-making requirements, therefore, should be
a key moderator of the effects of work-group diversity on
performance.
Although diversity research typically has not focused on task
requirements, there is some evidence in support of this proposition.
Jehn et al. (1999) found that informational diversity was positively
related to performance when task complexity was high rather than
low. The meta-analysis by Bowers et al. (2000) similarly suggests
that diverse groups may outperform homogeneous groups on dif-
ficult tasks, whereas homogeneous groups might in fact outper-
form diverse groups on more simple tasks.
Proposition 2: Task requirements moderate the relationship
between diversity and performance such that diversity may be
positively related to performance when performance requires
information processing and creative, innovative solutions.
The CEM’s emphasis on elaboration also points to task moti-
vation and ability as core moderators of the positive effects of
diversity. Motivation and ability are more or less neglected vari-
ables in diversity research, but they are seen as the primary
moderators of deep-level processing of information in social psy-
chological models of social perception and judgment (Chaiken &
Trope, 1999). The proposition that deep-level processing of infor-
mation is contingent on both the motivation and the ability to
conduct such processing is supported in research in a variety of
domains, including persuasive communication (Eagly & Chaiken,
1993), person perception (Fiske, 1998), negotiation (De Dreu &
Carnevale, 2003), and group decision making (Kruglanski & Web-
ster, 1991). Accordingly, we propose that motivation and ability
moderate the relationship between diversity and performance.
In the area of diversity, there is some support for this proposed
role of motivation in a recent experiment by Scholten, van Knip-
penberg, Nijstad, and De Dreu (2003). Scholten et al. focused on
the role of accountability in informationally diverse groups. Ac-
countability to others for the way in which decisions are made has
been shown to motivate more elaborate processing of information
(Lerner & Tetlock, 1999). The proposed role of motivation there-
fore suggests that accountability, as a factor affecting motivation,
should influence performance in diverse groups. Scholten et al.
examined the effects of accountability on group decision quality in
a decision task in which the best decision alternative could be
identified through the exchange and processing of information
uniquely assigned to individual group members (see Stasser,
1999). As predicted, and consistent with the proposed role of
motivation, Scholten et al. found that accountable groups engaged
in deeper information processing and reached better quality
decisions.
Proposition 3: Diversity is more likely to engender elabora-
tion and to benefit performance when group member task
motivation is high rather than low.
Proposition 4: Diversity is more likely to engender elabora-
tion and to benefit performance when group member task
ability is high rather than low.
Research Implications of Propositions 1 Through 4
Propositions 1 through 4 point to a number of factors that have
received little if any attention in diversity research. Proposition 1
points to a first reason why diversity research yields inconsistent
findings. Elaboration is likely to be more unambiguously related to
task performance than conflict and dissent, and a focus on elabo-
ration and the conditions under which diversity engenders elabo-
ration (i.e., Propositions 2– 4) is therefore likely to yield more
consistent relationships with performance. An obvious implication
for diversity research would therefore be to assess elaboration
processes and to focus on the contingencies of elaboration.
Because the focus is on group-level elaboration (Hinsz et al.,
1997), elaboration may in part be inferred from behavioral data
that may for instance be obtained through audio–video recording
of task groups. Such data can be coded for the exchange of
information and perspectives (Stasser, 1999) and for discussion
building on these exchanges. In addition, elaboration might be
inferred from questionnaire data assessing self-reported elabora-
tion, including the individual-level processing that may not be
1012
VAN KNIPPENBERG, DE DREU, AND HOMAN

inferred from behavioral observations and aggregating these data
to the group level (Homan & van Knippenberg, 2003). Alterna-
tively, research may be conducted along the lines of the study by
van Ginkel and van Knippenberg (2003) discussed above and
experimentally manipulate the extent to which groups engage in
elaboration of task-relevant information and perspectives.
Although there already is some evidence for the moderating role
of task requirements (Bowers et al., 2000; Jehn et al., 1999), future
research should provide more direct evidence for Proposition 2
regarding the role of task information processing, creativity, and
decision-making requirements. This may be done by including
measures of elaboration of task-relevant information and perspec-
tives in diversity research and relating these to either measures of
task characteristics (cf. McGrath, 1984; Stewart & Barrick, 2000)
or to experimental manipulations of task requirements.
Proposition 3 suggests that an important avenue for future
diversity research is to focus on motivation. One possibility is to
assess group members’ intrinsic motivation, to aggregate this to
the work-group level, and to test interactions with dimensions of
diversity. Because a host of factors may feed into motivation,
Proposition 3 also points to a range of factors that may moderate
the effects of diversity on performance through their effects on
motivation. Examples include leadership (e.g., D. van Knippen-
berg & Hogg, 2003), goal-setting (e.g., Locke & Latham, 1990),
organizational justice (e.g., Tyler, 1999), social exchange pro-
cesses (e.g., Rhoades & Eisenberger, 2002), individual differences
in information-processing motivation (e.g., De Dreu & Carnevale,
2003; Kruglanski & Webster, 1996), accountability (e.g., Lerner &
Tetlock, 1999), and group members’ motivation to work with the
group (e.g., Weaver, Bowers, Salas, & Cannon-Bowers, 1997).
Because these factors typically have not been studied in relation to
diversity, an important contribution of Proposition 3 is to bridge
diversity research and the broad range of theories focusing on task
motivation.
To directly test Proposition 4 about the moderating role of
ability one may assess, for instance, intelligence or level of edu-
cation (Hunter & Schmidt, 1996), aggregate this to the group level,
and test interactions between these proxies to ability and diversity.
In addition to general cognitive ability, task-specific knowledge,
skills, and abilities may be important, because they provide the
common ground and shared frame of reference that may aid
diverse groups in making sense of divergent information and
perspectives. Communication skills too may be an important as-
pect of a work group’s ability to process diverse information,
because part of the group-level elaboration process is the effective
communication of new information and perspectives (cf. the dis-
tinction between task skills and team skills; Kozlowski & Bell,
2003). Accordingly, a test of hypotheses derived from Proposition
4 might focus on task-specific knowledge, skills, and abilities;
communication skills; and more general measures of ability.
In a similar vein as Proposition 3, Proposition 4 points to a range
of factors that may influence ability as moderators of the relation-
ship between diversity and performance. The ability to elaborate
task-relevant information may be something that group members
bring to the group, but it may also be affected by situational
factors. Elaborate processing of task-relevant information and per-
spectives requires time (cf. Schweiger, Sandberg, & Rechner,
1989). Time constraints (e.g., deadlines) may thus put serious
limits on a group’s ability to elaborate. An experimental study by
Kruglanski and Webster (1991) provides some evidence in support
of this hypothesis. They observed that groups that worked on a
decision task under time pressure were less open to divergent
perspectives than groups that worked without such time con-
straints. Constraints on ability may also come from the working
environment. Environmental “noise” may interfere with a group’s
ability to exchange and process diverse information. Kruglanski
and Webster’s (1991) second experiment provides evidence for
this: Groups working in a noisy room were less open to divergent
perspectives than groups not exposed to noise.
In relation to ability, we may also note that even highly skilled
and accomplished group members may reach a point where their
ability is not up to par with group diversity. For a given level of
ability, we may assume that a group is able to benefit from
diversity up to a certain degree of diversity but that diversity
beyond this level will have no further benefit and may even disrupt
group process with increasing diversity by introducing misunder-
standings and so forth. Thus, even though more capable groups
may be better equipped to harvest the benefits of diversity, the
relationship of diversity with elaboration and performance is likely
to be curvilinear rather than linear, with diversity stimulating
elaboration and enhancing performance (contingent on ability) up
to a point, beyond which more diversity no longer benefits per-
formance and might even be detrimental to performance (i.e., an
inverted U-shape). In contrast to earlier work (cf. Williams &
O’Reilly, 1998), the CEM thus suggests that diversity can have
both positive and negative effects on performance through
information/decision-making processes. Although more research
is certainly needed, some early evidence for this derives from a
study by Gonzalez-Roma, West, and Borrill (2003), who found
curvilinear relationships with team innovativeness for team diver-
sity in gender, age, and tenure. Brodbeck (2003), who distin-
guished between business game teams with low, moderate, and
high diversity, likewise found that moderately diverse teams out-
performed teams with lower and higher diversity.
Diversity, Social Categorization, and Intergroup Bias
The information/decision-making perspective and the social cat-
egorization perspective have largely developed along separate
lines. Accordingly, little theoretical or empirical work in diversity
research concerns the question of how social categorization pro-
cesses may affect information/decision-making processes and,
more specifically, elaboration processes. This is a question of
considerable importance, because, as we argue in the next section,
there is reason to expect that social categorization processes may
disrupt elaboration of task-relevant information. Before we address
this issue, however, we first consider the relationships among
diversity, social categorization, and intergroup bias, because it is
important to distinguish among the determinants of social catego-
rization per se and the determinants of intergroup bias in research
on diversity and performance.
Most empirical research from the social categorization perspec-
tive on diversity has been conducted under the implicit assumption
that social categorization is merely a function of the degree of
difference between group members. Moreover, such research has
generally adopted an independent effects approach, assuming that
diversity on a dimension (e.g., ethnicity) is predictive of social
categorization independent from diversity on other dimensions
(e.g., age; Pelled et al., 1999). An implicit assumption operating in
such research is that social categorization more or less inevitably
1013
SPECIAL SECTION: WORK GROUP DIVERSITY

results in intergroup biases, that is, less positive reactions to others
categorized as dissimilar, or of out-group status, than to others
categorized as similar, or of in-group status (cf. Williams &
O’Reilly, 1998). Accordingly, most field studies of diversity use a
measure of diversity that reflects the degree of similarity among
group members on a given dimension and relate this to outcomes
without assessing mediating processes (e.g., Riordan & Shore,
1997; Tsui et al., 1992). In a similar vein, experimental studies
typically do not go beyond manipulating group composition as
being either diverse or homogeneous on a dimension of interest
(e.g., Cox et al., 1991). Interactions between different dimensions
of diversity typically are not tested, either in the field or in the lab
(Pelled et al., 1999; Williams & O’Reilly, 1998).
These common practices are unfortunate because research and
theory on social categorization, social identity, and intergroup
relations suggests more complex relationships between diversity
and categorization and between categorization and intergroup bias
(e.g., Brewer & Brown, 1998; Brown & Gaertner, 2001; D. van
Knippenberg, 2003). Accordingly, applying insights acquired from
diversity research should improve the explanatory power of mod-
els of work-group diversity.
The Salience of Social Categorizations
First, we address the relationship between diversity and social
categorization. Diversity may lead group members to distinguish
between “us” and “them.” Many work groups offer several poten-
tial bases for such us-versus-them distinctions (e.g., men vs.
women, old vs. young, sales vs. production), and not all of these
potential categorizations are necessarily used by the individuals in
the particular situation in question. A core question for diversity
research, therefore, is which factors determine the likelihood that
individuals view the self and others in terms of a particular
categorization, that is, what renders it salient in people’s minds?
The answer to this question is important to our understanding of
diversity effects, because it directly relates to the likelihood that
differences between work-group members elicit social categoriza-
tion processes. As a case in point, Randel (2002) assessed the
salience of gender categorization in a questionnaire and found that
salience moderated the relationship between gender composition
and relational conflict such that the relationship was stronger when
gender was more salient. Somewhat surprisingly, then, diversity
research typically has not focused on the determinants of social
category salience.
Self-categorization theory proposes that the salience of social
categorizations is a function of three factors (Oakes, Haslam, &
Turner, 1994; Turner et al., 1987): comparative fit, normative fit,
and cognitive accessibility. The comparative fit of a categorization
refers to the extent to which the categorization provides a good
reflection of similarities and differences between people. The more
a categorization results in high within-category similarity and high
between-category differences, the higher is its comparative fit and
the more likely the categorization is to be salient. Consider, for
example, transnational teams, in which members come from sev-
eral different countries. Consistent with the principle of compar-
ative fit, Earley and Mosakowski (2000) argued that within a team
consisting of many different nationalities, categorization into
same-nationality subgroups would not be easy, if at all possible,
whereas within teams consisting of two nationalities, categoriza-
tion in terms of nationality would be more salient (also see Lau &
Murnighan, 1998). These authors indeed observed more subcat-
egorization on perceptual–attitudinal measures within teams incor-
porating two nationalities than within teams incorporating several
different nationalities or within those that are homogeneous. The
effects of comparative fit have mainly been demonstrated in re-
search on cross-categorization, which points to the fact that dif-
ferent dimensions of diversity may interact in determining the
salience of social categorizations (Brewer, 1995; Lau & Mur-
nighan, 1998). When people differ on more than one dimension,
differences may either converge (i.e., be correlated) or cross-cut
each other (i.e., be unrelated). When differences converge (e.g., all
the male members of a work group are relatively young while all
the female members are relatively old), social categorization (i.e.,
younger men vs. older women) is more likely then when differ-
ences along the dimensions cross-cut each other (e.g., older and
younger group members are equally likely to be either male or
female; Homan & van Knippenberg, 2003; Marcus-Newhall,
Miller, Holtz, & Brewer, 1993; Oakes, Turner, & Haslam, 1991).
The second component underlying the salience of social cate-
gorization, namely, normative fit, reflects the extent to which the
categorization makes sense in relation to the individual’s cognitive
frame of reference (e.g., beliefs, expectations, stereotypes). The
more an individual believes that within the given context differ-
ences along a certain dimension are meaningful, the more salient
the categorization based on these differences will be. Stangor,
Lynch, Duan, and Glass (1992), for example, showed that indi-
viduals with racial prejudices (i.e., for whom race is more strongly
associated with meaningful differences) are more likely to use race
categorizations than are those who are less racially prejudiced. The
subjective meaningfulness of categorization depends not only on
individual differences but also on contextual differences. Some
jobs, for instance, such as construction work or truck driving, seem
to be more strongly associated with stereotypic beliefs about
gender differences than others. In line with this view, D. van
Knippenberg, Haslam, and Platow (2003) proposed that the extent
to which group identification is affected by diversity on a partic-
ular dimension is contingent on beliefs about the extent to which
that dimension is relevant to the task at hand (i.e., has normative
fit). To test this hypothesis experimentally, D. van Knippenberg et
al. gave participants bogus feedback on two alleged personality
dimensions about the composition of a group to which they were
assigned. One dimension was indicated as related to performance
on the group task and the other dimension, to performance on a
task the participants would perform individually later in the ex-
periment. As predicted, group identification was only affected by
group composition on the dimension indicated as relevant to group
performance and not by the group composition on the other di-
mension. These results suggested that it is not difference per se but
rather the belief that a difference is meaningful within the context
of the group that leads diversity to affect categorization processes.
The final component underlying the salience of social catego-
rization, cognitive accessibility of a categorization, refers to the
ease with which the categorization comes to mind and the readi-
ness of the perceiver to use the categorization. Accessibility de-
pends on prior experience, beliefs, and expectations but also on
contextual cues that may “prime” a categorization. Well-learned
social categorizations that people have been socialized to use
throughout their lives, such as those based on sex and age, should,
for instance, be more accessible than nonobvious categorizations,
such as those based on the color of people’s cars or on toothpaste
1014
VAN KNIPPENBERG, DE DREU, AND HOMAN

preference (Fiske, 1998; Stangor et al., 1992). Similarly, contex-
tual cues such as a team supervisor who continually refers to
functional differences may, for example, make a categorization
based on functional background more accessible (Macrae, Boden-
hausen, & Milne, 1995; A. van Knippenberg & Dijksterhuis,
2000).
Self-categorization theory proposes that the salience of a social
categorization requires all three components just discussed and
thus suggests that these components interact (Turner et al., 1987):
The higher comparative fit, normative fit, and cognitive accessi-
bility, the more likely a categorization is to be salient. Although it
should be noted that to our knowledge, no study has yet tested the
full three-way interaction, in support of this proposition, reviews
of the social categorization literature have yielded consistent evi-
dence for the two-way interactions implied in this three-way
interaction (Oakes et al., 1994; A. van Knippenberg & Dijkster-
huis, 2000). In sum, then, we propose that differences among
group members do not necessarily result in social categorization,
nor do greater differences automatically result in stronger catego-
rization. The extent to which they do so is contingent on perceiv-
ers’ readiness to use the categorization, the extent to which the
categorization is subjectively meaningful, and the extent to which
the categorization results in relatively homogeneous categories
that are clearly different from each other.
Proposition 5: Social categorization within work groups
is contingent on the interaction of the comparative fit, the
normative fit, and the cognitive accessibility of social
categorizations.
Proposition 5 suggests that another reason why diversity re-
search has yielded inconsistent results may be that the contingen-
cies of social category salience typically have not been taken into
account, and the clear implication for diversity research, therefore,
is to develop research in this area, as we discuss below. There is,
however, a second major consideration with regard to work-group
diversity and social categorization processes, which is that social
categorization should not be equated with intergroup biases.
The Relationship Between Social Categorization and
Intergroup Bias
Even though research in intergroup relations identifies social
categorization as the root cause of problematic intergroup attitudes
and behavior, social categorization should not be equated with
intergroup bias. Intergroup bias refers to more favorable percep-
tions of, and attitudes and behavior toward, in-group than out-
group (Brewer, 1979), whereas social categorization merely refers
to the perceptual grouping of people (Turner et al., 1987). The
distinction between social categorization and intergroup bias is
important for diversity research, because the potentially negative
effects of diversity identified in the social categorization perspec-
tive are clearly linked to intergroup bias and not to social catego-
rization per se. Intergroup biases negatively influence affective–
evaluative reactions to diverse groups and their members. In effect,
such biases may be expected to translate into the outcomes pre-
dicted in the social categorization perspective: low interpersonal
liking and group cohesion, relational conflicts, low identification
with and commitment to the group, low satisfaction, and high
turnover (intentions), to name but a few (Williams & O’Reilly,
1998). Diversity research has more or less ignored the distinction
between categorization and bias and seems to proceed on the basis
of the implicit assumption that bias is the inevitable consequence
of categorization. Growing evidence suggests this assumption to
be incorrect (Brewer & Brown, 1998; Brown & Gaertner, 2001; D.
van Knippenberg, 2003), and thus points to the value of distin-
guishing between factors underlying categorization per se (i.e.,
Proposition 5) and factors underlying intergroup biases in diversity
research.
Group memberships reflect on how individuals see the self (i.e.,
social identity). Accordingly, because individuals value a positive
and distinctive self-image, group members value a positive and
distinctive group identity (Brewer, 1991; Hogg & Abrams, 1988;
Tajfel & Turner, 1986). As a result, intergroup biases are typically
inspired by threats or challenges to the value or the distinctiveness
of group identity (i.e., threats to individuals’ self-views), whereas
social categorization is less likely to result in intergroup biases in
the absence of such threats or challenges (Branscombe, Ellemers,
Spears, & Doosje, 1999). Indeed, in the absence of such threats
and challenges, diversity may sometimes be valued more than
homogeneity (D. van Knippenberg & Haslam, 2003). Threats to
the value of identity may take many different forms but all share
that a favorable image of the group is challenged. Such challenges
may range from subtle social competition for status and prestige to
outright derogation or discrimination of the group (e.g., Brewer &
Brown, 1998) and include unequal status of subgroups and com-
petitive interdependence between subgroups (Gaertner & Dovidio,
2000). Threats to distinctiveness may be experienced when the
identity of the group as different from other groups is denied or
suppressed, for instance, when management highlights a shared
superordinate identity (e.g., organizational, work group) with an
emphasis on within-group similarity that downplays differences
between subgroups (Hornsey & Hogg, 2000). This observation
implies that trying to suppress categorizations may sometimes
result in stronger intergroup biases than explicitly recognizing
these categorizations in addition to recognizing the superordinate
categorization as a work group (Hewstone & Brown, 1986; Horn-
sey & Hogg, 2000; van Leeuwen & van Knippenberg, 2003; see
also Gaertner & Dovidio, 2000). As an illustration, consider
Dovidio, Gaertner, and Validzic’s (1998) study of intergroup bias
in diverse groups. These authors found that intergroup biases were
lower when subgroups had distinct roles (reducing distinctiveness
threat) and equal status (reducing threats to value) and, at the same
time, held more inclusive (superordinate) perceptual representa-
tions of the group.
Proposition 6: Social categorization results in intergroup bi-
ases that are disruptive to group functioning to the extent that
the identity implied by the categorization is subjectively
threatened or challenged.
Research Implications of Propositions 5 and 6
Proposition 5 points to the need to study the contingencies of
social category salience in diversity research. Thus far, applica-
tions of the principles of the salience of social categorizations in
diversity research are scarce and mainly concern comparative fit
(e.g., Earley & Mosakowski, 2000; Lau & Murnighan, 1998). The
principle of comparative fit suggests that different dimensions of
diversity interact in eliciting social categorization processes. Part
1015
SPECIAL SECTION: WORK GROUP DIVERSITY

of the inconsistent findings in diversity research can thus be
attributed to the independent effects approach (cf. Pelled et al.,
1999), and diversity research would do well to shift to an interac-
tive effects approach. Note, however, that testing interactions
between dimensions of diversity by itself is insufficient, because a
hypothesis about these interactive effects requires information
about the extent to which the dimensions are correlated (i.e., to
what extent do they converge or cross-cut each other?). This was
nicely illustrated by Pelled et al., who obtained a number of
interactions among dimensions of diversity. The authors found that
the interactions had either positive or negative effects contingent
on the correlation between the dimensions. The interactive effects
of different dimensions of diversity may also be studied experi-
mentally, bringing the correlation among dimensions under exper-
imental control (e.g., Homan & van Knippenberg, 2003; see be-
low). Another option is the approach advanced by Thatcher, Jehn,
and Zanutto (2003), who suggested replacing or complementing
the similarity indices of diversity that are typically used in diver-
sity research (i.e., reflecting similarity on a single dimension) by a
measure that is based on similarities and differences on multiple
dimensions simultaneously and that reflects the extent to which
these differences converge or cross-cut each other.
The principle of normative fit suggests that in order to under-
stand the relationship between diversity and social categorization,
researchers should study not only group composition as such but
also stereotypes and diversity beliefs, attitudes, and culture (Chat-
man, Polzer, Barsade, & Neale, 1998; D. van Knippenberg &
Haslam, 2003). As the D. van Knippenberg et al. (2003) study
suggested, however, these aspects should not necessarily be seen
as general beliefs or attitudes about diversity but as more context-
specific. A diversity dimension that may be subjectively relevant
in one context need not be so in another. This suggests that another
reason for the inconsistent findings in diversity research is that a
dimension of diversity may be subjectively meaningful in the
context of one study but not in the context of another. A challenge
to diversity research, therefore, is to develop theory to account for
and predict the normative fit of dimensions of diversity. In the
present context, an obvious starting point would be the effects that
task requirements may have on normative fit (cf. D. van Knippen-
berg et al., 2003).
The role of cognitive accessibility, as of yet, has hardly been
explored in diversity research. For some of the more frequently
studied dimensions of diversity, such as gender, age, and ethnicity,
accessibility might not be an issue, because such dimensions may
be chronically accessible (Stangor et al., 1992; A. van Knippen-
berg & Dijksterhuis, 2000). For other dimensions of diversity,
however, such as functional differences, accessibility may be more
of an issue, and it is worthwhile to explore how individual differ-
ences and contextual cues influence the accessibility of such
categorizations in diverse work groups.
Proposition 6 suggests that because social categorization does
not necessarily result in intergroup bias, diversity research should
study intergroup bias in addition to social categorization. One of
the things this implies is assessing rather than assuming social
categorization and intergroup bias (for examples, see Gaertner &
Dovidio, 2000). Equally important is a focus on threats and chal-
lenges to identity as moderators of the relationship between social
categorization and intergroup biases. This would involve assessing
the role of such factors as subgroup status (Gaertner & Dovidio,
2000; Hornsey & Hogg, 2000), cooperative versus competitive
intergroup relations (Gaertner & Dovidio, 2000), and distinctive
subgroup roles or functions (Hewstone & Brown, 1986; van Leeu-
wen & van Knippenberg, 2003).
Although recognizing differences may be important under some
circumstances, individuals may also be categorized against their
will (e.g., asking for the “female perspective” when this is seen as
inappropriate; Moss Kanter, 1977). Such categorization threats
(Branscombe et al., 1999) may invite individuals to distance them-
selves from the situation and the imposed categorization (Brewer,
Manzi, & Shaw, 1993; van Prooijen & van Knippenberg, 2000).
This poses a challenge from a more applied perspective, because it
suggests that neither acknowledging nor downplaying categoriza-
tions has uniformly positive or negative effects. Diversity research
thus should develop theory to predict the contingencies of distinc-
tiveness threat and categorization threat in diverse work groups.
Resolving this issue would greatly help diversity management in
revealing under what circumstances diversity, and social catego-
rizations, should be made an issue.
Interactive Effects of Elaboration and Social
Categorization Processes
Intergroup Bias May Disrupt Elaboration
Propositions 1 through 4 concern elaboration as the core process
underlying the positive effects of diversity, and Propositions 5 and
6 offer a more refined look at social categorization processes as
potentially disruptive to diverse groups. As such, the propositions
all point to potential reasons why research on diversity and per-
formance has yielded such inconsistent results. In this section, we
focus on yet another reason for the inconsistency in findings,
which is the interaction between elaboration processes and social
categorization processes in the relationship between work-group
diversity and group performance.
As we noted in the introduction, the information/decision-
making and social categorization perspectives have largely devel-
oped in separate traditions (Williams & O’Reilly, 1998). Our
analysis thus far has likewise considered these perspectives in
isolation. Propositions 1 through 4 reflect a reconceptualization of
information/decision-making processes, and Propositions 5 and 6
reflect a refinement of our thinking about the salience of social
categorization and the not-so-inevitable link between social cate-
gorization and intergroup biases. In this section we integrate these
two separate streams of thought and thus develop the CEM in full.
Research in persuasive communication has identified intergroup
biases as a factor disrupting elaboration of others’ communica-
tions. Intergroup biases lead people to place more trust in in-group
than in out-group members and to see in-group members as more
valid sources of information (Brewer, 1979; Turner et al., 1987).
As a result, communications from in-group members are more
likely to be attended to and elaborated, and thus more likely to
influence the thoughts and actions of the individual, than are
communications from out-group members (Mackie, Worth, &
Asuncion, 1990; D. van Knippenberg, 1999; D. van Knippenberg
& Wilke, 1992; see also Clark & Maass, 1988). Because the
elaboration of task-relevant information that is proposed to under-
lie the positive effects of work-group diversity on performance
(Proposition 1) requires attention to and elaboration of communi-
cations from diverse others (i.e., as sources of novel information
and perspectives), we propose that intergroup biases elicited by
1016
VAN KNIPPENBERG, DE DREU, AND HOMAN

diversity interfere with diversity’s potential to elicit elaboration of
task-relevant information.
Consider, for example, a merger of two banks (cf. Buono,
Bowditch, & Lewis, 1985). Within newly formed teams providing
financial services, a distinction between “us from our bank” and
“them from the other bank” is easily made. This distinction, in and
of itself, is not problematic. Intergroup biases, however, may lead
team members to view the way of working that was customary at
their former organization as superior to that of the other bank and
to see team members from the other bank as less expert. These
intergroup biases may also render team members less willing to
invest in outgroup others and to keep them fully informed and
up-to-date in matters concerning the team. As a result, elaboration
of task-relevant information suffers, because team members ex-
change insufficient information and pay inadequate attention to
perspectives and information introduced by outgroup others when
preparing and developing financial services, and the synergy often
hoped for in mergers is not achieved.
Proposition 7: Intergroup biases elicited by work-group di-
versity are disruptive to elaboration of task-relevant informa-
tion and therefore to group performance.
Because diversity research typically has not focused on both
information/decision-making and social categorization simulta-
neously, evidence from diversity research pertaining to our pro-
posed integration of the two perspectives is rare. Bhappu, Griffith,
and Northcraft (1997) compared mixed-gender groups with
gender-homogeneous groups and found, in line with Proposition 7,
that diversity biased attention and influence in communication.
Jehn et al. (1999) found that both social category diversity and
value diversity were disruptive to positive relationships of infor-
mational diversity with performance indicators. It is unclear, how-
ever, whether the results of the Jehn et al. study should be attrib-
uted to the disruptive effects of intergroup biases elicited by social
category and value diversity or to interactive effects with infor-
mational diversity on category salience (cf. comparative fit), or to
both factors. Less ambiguous evidence in this respect comes from
a recent experiment by Homan and van Knippenberg (2003) in
which a manipulation of gender composition was crossed with a
manipulation of informational diversity (using different sets of
background information) to yield, amongst others, a condition in
which informational diversity cross-cut gender diversity and a
condition in which dimensions of diversity converged. Informa-
tional diversity was expected to elicit more elaboration (assessed
in questionnaire data aggregated to the group level) but only if
informational and gender diversity did not converge (cf. Proposi-
tion 5). Results supported this prediction and also showed that this
pattern of results was complemented by higher interpersonal liking
(as an indicator of lower intergroup bias) in the cross-
categorization condition than in the other gender-diverse groups.
These findings show that when diversity engenders categorization
processes resulting in intergroup bias, elaboration of task-relevant
information is disrupted.
Research Implications of Proposition 7
Thus, there is some evidence in support of Proposition 7 as well
as evidence suggesting that the contingencies of social categori-
zation identified in Proposition 5 moderate the extent to which
diversity may give rise to intergroup biases that disrupt elaboration
processes. Clearly, however, more research is needed, for instance,
on the role of other contingencies of social categorization identi-
fied in Propositions 5 and the contingencies of intergroup bias
identified in Proposition 6. In addition, an important implication of
Proposition 7 is that when studying information/decision-making
(i.e., elaboration) processes, one should assess social categoriza-
tion and intergroup bias as well as elaboration and take the con-
tingencies of social categorization and intergroup bias into con-
sideration. Proposition 7 also suggests that one reason why
diversity research has been rather unsuccessful in consistently
linking diversity to performance is that the information/decision-
making and social categorization perspectives have largely devel-
oped along separate lines, whereas categorization processes need
to be taken into account to reliably predict elaboration processes.
Categorization and Elaboration Are Not Tied to Specific
Dimensions of Diversity
Social Category Diversity and Informational Diversity as
Different Aspects of Diversity Dimensions
Our analysis thus far has shown how a different and sometimes
more refined look at the nature of information/decision-making
processes and social categorization processes can help explain the
inconsistent findings in diversity research. An issue we have not
yet dealt with but that is also addressed by the CEM is whether
some types of diversity are more or less likely to be associated
with some of these processes (e.g., social categorization) than
others. Past work tends to make a distinction between informa-
tional (functional, job-related) diversity and social category (de-
mographic, not job-related) diversity. Information/decision-
making processes have been linked to informational diversity,
whereas social categorization processes have been associated with
social category diversity (e.g., Jehn et al., 1999; Pelled et al.,
1999). Qualitative (Milliken & Martins, 1996; Williams &
O’Reilly, 1998) and quantitative (Bowers et al., 2000; Webber &
Donahue, 2001) reviews fail, however, to support this proposed
association of diversity type with process and performance across
studies. We have identified a number of reasons for the lack of
empirical support. The CEM proposes yet another reason: the
distinction between types of diversity is flawed.
The distinction between social category diversity and informa-
tional diversity may be intuitively appealing, but it is not as
clear-cut as it may seem on first sight. It is not as easy as it might
appear to determine which differences are task-relevant. Some
dimensions of diversity are obviously task-relevant, such as dif-
ferences in educational and functional background. Differences
that appear not to be job-related at first sight, however, may also
be associated with task-relevant information and perspectives. In
developing personnel policy, for instance, married versus nonmar-
ried individuals, men versus women, or older versus younger
individuals may provide different task-relevant perspectives on
what is important to employees, even though marital status, gen-
der, and age are typically not seen as relevant criteria in selection
for personnel management. In other words, what are typically seen
as social category differences may be associated with task-relevant
differences (Tsui & O’Reilly, 1989). Social category differences
thus are confounded with informational differences and, as a
result, dimensions of diversity that are typically conceptualized as
1017
SPECIAL SECTION: WORK GROUP DIVERSITY

social category diversity (e.g., gender, ethnicity) may elicit the
positive effects implied in the information/decision-making per-
spective (e.g., Cox et al., 1991), that is, as we argue in the present
study, elaboration of task-relevant information and perspectives.
Conversely, informational differences may give rise to social
categorization processes (e.g., Homan & van Knippenberg, 2003;
Marcus-Newhall et al., 1993). Not only may informational differ-
ences be visible through, for instance, differences in dress, non-
visible differences also may be recognized, form the basis for
categorizations that seem meaningful, and be readily used within a
particular context (e.g., the economists vs. the psychologists in a
business school, opponents vs. proponents of a particular position).
Thus, what may appear to be clear informational differences may
in practice work as social category differences, acting as a basis for
categorization processes in their own right. We therefore propose
that informational diversity and social category diversity are as-
pects or functions of diversity rather than types of diversity, the
key point being that all dimensions of diversity may elicit catego-
rization processes as well as elaboration processes.
Proposition 8: All dimensions of diversity may elicit social
categorization processes as well as elaboration processes.
Research Implications of Proposition 8
An important implication of Proposition 8 is that if any dimen-
sion of diversity can function as both social category diversity and
informational diversity, any dimension can elicit both the effects
associated with elaboration processes and the effects associated
with categorization processes. Any model of diversity, then, that
assumes that categorization processes are tied to certain dimen-
sions of diversity while elaboration (cf. information/decision mak-
ing) processes are tied to other dimensions is bound to fall short of
reliably predicting the positive and negative performance effects of
diversity. The clear implication for diversity research therefore is
to abandon the focus on typologies of diversity to explain the
differential effects of diversity, and to focus on the contingencies
of elaboration, categorization, and intergroup bias to predict which
function of diversity will prevail in a given context.
Research on group tenure and diversity suggests that the func-
tion of dimensions of diversity may change over time. A number
of studies show that diverse groups may initially perform worse
than more homogeneous groups but, over time, may perform as
well as or even outperform homogeneous groups (Earley & Mo-
sakowski, 2000; Schippers, Den Hartog, Koopman, & Wienk,
2003; Watson, Kumar, & Michaelsen, 1993; see also Gruenfeld,
Mannix, Williams, & Neale, 1996). In addition, Harrison, Price,
and Bell (1998) found a negative relationship between demo-
graphic diversity and team integration (a concept akin to cohesion)
that diminished over time, whereas a negative relationship of
attitude diversity with integration only became apparent over time.
These moderating roles of tenure may be explained from Propo-
sition 8 and the contingencies of elaboration, categorization, and
intergroup bias.
First, group members may need time to develop the knowledge,
skills, and abilities required to elaborate divergent perspectives and
to develop a shared sense of “who knows what,” that is, shared
knowledge, referred to as transactive memory (Wegner, 1986), or
a shared mental model of the group (Cannon-Bowers & Salas,
2001). Transactive memory affects the team’s ability for elabora-
tion, because it helps determine which group members to consult
on particular issues and provides a frame of reference for placing
the individual’s contribution in context (Moreland, 1999). Second,
comparative fit may change as new similarities and differences are
uncovered through interaction (cf. Swann, Polzer, Seyle, & Ko,
2004). Similarly, normative fit may change as people get to know
dissimilar others better over time and as experience overrides
stereotype-based perceptions of similarities and differences
(Gaertner & Dovidio, 2000; Harrison et al., 1998). Extended
tenure may thus result in lower salience of the categorizations
initially engendered by diversity. In combination, this suggests a
shift over time from categorization processes to information/
decision-making processes, which explains the change in perfor-
mance of diverse groups relative to homogeneous groups observed
in diversity research. At the same time, however, as Harrison et
al.’s (1998) findings for attitude diversity suggest, over time
groups may also uncover more hidden differences that may give
rise to disruptive categorization processes. The conclusion should
therefore be that the processes elicited by diversity may change
over time rather than that there will always be a shift from
categorization to elaboration processes. The earlier studies of the
moderating role of group tenure were not designed to test the
proposed roles of elaboration, categorization, and intergroup bias,
however, and more process-oriented research on the moderating
role of tenure is clearly needed to substantiate this conclusion.
There seems to be an emerging interest in other dimensions of
diversity, such as diversity in personality (e.g., Barrick, Stewart,
Neubert, & Mount, 1998; Van Vianen & De Dreu, 2001), attitudes
(e.g., Harrison et al., 1998), and values (e.g., Jehn et al., 1999). An
obvious question from the perspective of the current discussion is
to what extent Proposition 8 also applies to these dimensions. It
seems obvious that differences in attitudes and values may intro-
duce different perspectives as well as different information asso-
ciated with these perspectives. Personality differences too may be
associated with divergent perspectives and information, because an
individual’s personality is likely to shape the individual’s back-
ground, invite specific experiences, and so forth. At the same time,
differences on all of these dimensions may lead to the perception
that others are different from the self and invite social categoriza-
tion processes. Indeed, Harrison, Price, Gavin, and Florey (2002;
also see Harrison et al., 1998) found that diversity in personality,
beliefs, attitudes, and values was negatively related to team inte-
gration. Thus, it appears plausible that diversity in attitudes, val-
ues, and personality may elicit both elaboration and social cate-
gorization processes. Assessing these processes, and the
contingencies of these processes, would be a worthwhile avenue
for future research in this area.
Implications for Diversity Management
Although the primary aim of our analysis is to address the
inconsistent research findings for diversity and performance, and
to outline the implications of our model for diversity research, the
CEM has a number of implications for the management of diver-
sity. In short, the model suggests that to harvest the potential
benefits of work-group diversity, and to prevent its potentially
disruptive effects, diversity management should focus on fostering
elaboration (Proposition 1) and on preventing intergroup biases
that are associated with negative affective– evaluative reactions to
the group and its members (Proposition 6) and that may disrupt
1018
VAN KNIPPENBERG, DE DREU, AND HOMAN

elaboration of task-relevant information (Proposition 7). The mod-
erators of elaboration (Proposition 2, 3, and 4), categorization
(Proposition 5), and intergroup bias (Proposition 6) we identified
provide guidelines for management in these respects.
Proposition 2, for instance, implies that team composition de-
cisions may favor a diverse composition to the extent that task
requirements are conducive to the positive effects of diversity. The
proposition also suggests that diversity management should work
from the realization that task requirements impose important con-
straints on the extent to which positive effects of diversity on
group performance may be expected. For tasks that have strong
information-processing and decision-making components, the
CEM implies that diversity management should focus on task
motivation (Proposition 3) and task ability (Proposition 4). Moti-
vation, for instance, may be fostered by leadership that builds
intrinsic motivation (e.g., D. van Knippenberg & Hogg, 2003), by
treating group members fairly and respectfully (Tyler, 1999), by
providing organizational and supervisor support (Rhoades &
Eisenberger, 2002), and by setting challenging goals (Locke &
Latham, 1990). Ability may in part be managed through selection,
that is, by selecting individuals high in cognitive ability, task-
specific knowledge, skills, and abilities, and communication skills
into the work group. When it comes to selection and work-group
composition, our analysis of the moderating role of ability also
suggests that groups can become too diverse. To the extent that this
occurs at management’s discretion, then, differences should be
sought in moderation. Group members’ ability to elaborate diverse
information may also develop over time as members become more
familiar with each others’ perspectives and develop transactive
memory. This suggests that especially for diverse work groups it is
important that they can reach more extended tenure and that they
be allowed a more extended start-up phase than more homoge-
neous groups. Alternatively, training or leadership to recognize
differences in expertise and perspective may to a certain extent
substitute for extended tenure (Moreland, 1999; also see Stasser,
Vaughan, & Stewart, 2000).
Propositions 5 and 6 suggest that preventing disruptive inter-
group biases begins with the process of managing the salience of
social categories. Because it is unrealistic to assume that diversity
management successfully lowers the salience of all social catego-
rizations in each and every work group, diversity management
should also focus on managing the contingencies of intergroup
biases. Notions of social category salience suggest, for instance,
that comparative fit more than difference per se contributes to
social categorization salience. Eliminating diversity may be neither
feasible nor desirable, but preventing high comparative fit in
work-group composition should be a more realistic ambition,
because it does not require lowering diversity (cf. cross-
categorization). Team and work-group composition could thus
focus on cross-cutting dimensions of diversity, both for desired
(e.g., functional) and “natural” (e.g., demographic) dimensions of
diversity. Work by D. van Knippenberg et al. (2003; D. van
Knippenberg & Haslam, 2003) suggests that even when categori-
zations have normative fit and are salient, negative responses to
diversity need not result. These studies show that when differences
are seen as valuable to group functioning, group members may
respond more positively to diverse groups than to more homoge-
neous groups. Thus, part of the process of managing diversity
could also involve fostering such value-in-diversity beliefs and
attitudes (Ely & Thomas, 2001; D. van Knippenberg & Haslam,
2003). A focus on the value of diversity for the work group may
also build mutual respect and address potential threats to the value
and distinctiveness of identities (cf. Proposition 6).
Proposition 8 implies that the above-mentioned guidelines es-
sentially hold for all dimensions of diversity. Diversity manage-
ment therefore should not rely on typologies of diversity that imply
that only certain dimensions of diversity may yield positive effects,
whereas other dimensions of diversity may at best not be disrup-
tive, but instead work from the assumption that all dimensions of
diversity can stimulate performance through information elabora-
tion as well as hinder performance by causing disruptive inter-
group biases.
Conclusion: Setting the Research Agenda
The starting point for the current analysis was the inconsistency
in the various findings for the relationship between work-group
diversity and performance. To address this problem, we advanced
the CEM of work-group diversity and group performance. The
CEM centers around eight propositions, each in its own right
reflecting a different or more refined look at the way work-group
diversity affects—through information/decision-making (i.e., elab-
oration) processes and social categorization processes—work-
group performance. In addition to integrating past work on diver-
sity in such a way that many inconsistencies in research findings
can be explained, the CEM points to the need to study the pro-
cesses underlying the effects of work-group diversity. Because all
dimensions of diversity can, in principle, have both positive and
negative effects, processes cannot be assumed on the basis of the
dimension of diversity studied. Nor can they be assumed on the
basis of the relationship between diversity and performance: Social
categorization itself is not problematic, and elaboration processes
also can lead to negative effects (i.e., when group ability is not up
to par with group diversity). Thus, we echo recent calls for
process-oriented research on work-group diversity (Lawrence,
1997; Pelled et al., 1999), because such a focus will greatly
improve our understanding of the effects diversity can have.
Although our core propositions are consistent with laboratory
and field research on work-group diversity, many aspects of the
CEM require new research. For example, the CEM points to
moderating and mediating variables that have been ignored thus
far. Testing moderating relationships requires an increased focus
on statistical interaction effects instead of main effects. Because
statistical interactions are typically underestimated in survey re-
search (McClelland & Judd, 1993), proper assessment of the
proposed interactions requires complementing cross-sectional re-
search with controlled experiments. In the previous sections, we
have discussed several examples of what such experiments could
look like.
Our analysis focused on group performance more than on other
possible effects of diversity, such as turnover or commitment. One
reason is that the divergent perspectives on the positive versus the
negative effects of diversity primarily concern diversity’s effects
on group performance (i.e., the need for integration is greatest
here). A second reason is that diversity effects on other frequently
studied outcomes such as cohesion, commitment, and turnover
may be assumed to reflect in part the same processes that lead to
diversity’s effects on group performance and are indeed included
in the CEM (cf. Proposition 6). Ultimately, however, it should be
recognized that our discussion and our model do not concern all
1019
SPECIAL SECTION: WORK GROUP DIVERSITY

possible effects of diversity. We excluded discussion of such
issues as the relationship of diversity with the group’s ability to
maintain external relationships (Ancona & Caldwell, 1992), be-
cause these issues relate to a different set of processes than the
group dynamics involved in the more proximal effects of diversity,
and incorporation of these processes would go beyond the scope of
the present study. Exploring the implications of the CEM for other
outcomes than performance seems, however, an important avenue
for future research and theory development.
In conclusion, the present study provides a renewed and refined
look at the information-processing and social categorization pro-
cesses triggered by work-group diversity. In providing this per-
spective, we have integrated theoretical accounts of work-group
diversity, explained inconsistencies in past research findings, and
suggested several avenues for future research. We hope that the
combination of such research efforts will yield realistic advice
concerning how to manage effective performance in work groups
in an increasingly diverse society.
References
Ancona, D. G., & Caldwell, D. F. (1992). Demography and design:
Predictors of new product team performance. Organization Science, 3,
321–341.
Bantel, K., & Jackson, S. (1989). Top management and innovations in
banking: Does the composition of the team make a difference? Strategic
Management Journal, 10, 107–124.
Barrick, M. R., Stewart, G. L., Neubert, M. J., & Mount, M. K. (1998).
Relating member ability and personality of work-team processes and
team effectiveness. Journal of Applied Psychology, 83, 377–391.
Bhappu, A. D., Griffith, T. L., & Northcraft, G. B. (1997). Media effects
and communication bias in diverse groups. Organizational Behavior and
Human Decision Processes, 70, 199 –205.
Bowers, C., Pharmer, J. A., & Salas, E. (2000). When member homoge-
neity is needed in work teams: A meta-analysis. Small Group Research,
31, 305–327.
Branscombe, N. R., Ellemers, N., Spears, R. & Doosje, B. (1999). The
context and content of social identity threat. In N. Ellemers, R. Spears,
& B. Doosje (Eds.), Social identity: Context, commitment, content (pp.
35–58). Oxford, UK: Blackwell Science.
Brewer, M. B. (1979). In-group bias in the minimal intergroup situation: A
cognitive–motivational analysis. Psychological Bulletin, 86, 307–324.
Brewer, M. B. (1991). The social self: On being the same and different at
the same time. Personality and Social Psychology Bulletin, 17, 475–
482.
Brewer, M. B. (1995). Managing diversity: The role of social identities. In
S. E. Jackson & M. N. Ruderman (Eds.), Diversity in work teams (pp.
47– 68). Washington, DC: American Psychological Association.
Brewer, M. B., & Brown, R. J. (1998). Intergroup relations. In D. T. Gilbert
& S. T. Fiske (Eds.), Handbook of social psychology, (4th ed., pp.
554 –594). Boston: McGraw-Hill.
Brewer, M. B, Manzi, J. M., & Shaw, J. S. (1993). In-group identification
as a function of depersonalization, distinctiveness, and status. Psycho-
logical Science, 4, 88–92.
Brodbeck, F. (2003, May). Contradiction as an inhibitor and facilitator of
group performance. Paper presented at the 11th European Congress of
Work and Organizational Psychology, Lisbon, Portugal.
Brown, R., & Gaertner, S. L. (2001). Intergroup processes. Malden, MA:
Blackwell.
Buono, A. F., Bowditch, J. L., & Lewis, J. W. (1985). When cultures
collide: The anatomy of a merger. Human Relations, 38, 477–500.
Cannon-Bowers, J. A., & Salas, E. (2001). Reflections on shared cognition.
Journal of Organizational Behavior, 22, 195–202.
Chaiken, S., & Trope, Y. (1999). Dual process theories in social psychol-
ogy. New York: Guilford Press.
Chatman, J. A., Polzer, J. T., Barsade, S. G., & Neale, M. A. (1998). Being
different yet feeling similar: The influence of demographic composition
and organizational culture on work group processes and outcomes.
Administrative Science Quarterly, 43, 749 –780.
Clark, R. D., & Maass, A. (1988). Social categorization in minority
influence: The case of homosexuality. European Journal of Social
Psychology, 18, 347–364.
Cox, T. H., Lobel, S. A., & McLeod, P. L. (1991). Effects of ethnic group
cultural differences on cooperative and competitive behavior on a group
task. Academy of Management Journal, 34, 827– 847.
De Dreu, C. K. W., & Carnevale, P. J. (2003). Motivational bases of
information processing and strategy in conflict negotiation. Advances in
Experimental Social Psychology, 35, 235–291.
De Dreu, C. K. W., Harinck, S., & van Vianen, A. E. M. (1999). Conflict
and performance in groups and organizations. In C. L. Cooper & I. T.
Robertson (Eds.), International review of industrial and organizational
psychology (Vol. 14, pp. 369 – 414). Chichester, United Kingdom:
Wiley.
De Dreu, C. K. W., & Weingart, L. R. (2003a). A contingency theory of
task conflict and performance in groups and organizational teams. In M.
West, D. Tjosvold, & K. G. Smith (Eds.), International handbook of
organizational teamwork and cooperative working (pp. 151–166).
Chichester, United Kingdom: Wiley.
De Dreu, C. K. W., & Weingart, L. R. (2003b). Task and relationship
conflict, team performance, and team member satisfaction: A meta-
analysis. Journal of Applied Psychology, 88, 741–749.
De Dreu, C. K. W., & West, M. A. (2001). Minority dissent and team
innovation: The importance of participation in decision making. Journal
of Applied Psychology, 86, 1191–1201.
Dovidio, J. F., Gaertner, S. L., & Validzic, A. (1998). Intergroup bias:
Status, differentiation, and a common in-group identity. Journal of
Personality and Social Psychology, 75, 109 –120.
Eagly, A. H., & Chaiken, S. (1993). The psychology of attitudes. Orlando,
FL: Harcourt Brace Jovanovich.
Earley, P. C., & Mosakowski, E. (2000). Creating hybrid team cultures: An
empirical test of transnational team functioning. Academy of Manage-
ment Journal, 43, 26–49.
Ely, R. J., & Thomas, D. A. (2001). Cultural diversity at work: The effects
of diversity perspectives on work group processes and outcomes. Ad-
ministrative Science Quarterly, 46, 229 –273.
Fiske, S. T. (1998). Stereotyping, prejudice, and discrimination. In D. T.
Gilbert, S. T. Fiske, & G. Lindzey (Eds.), Handbook of social psychol-
ogy (4th ed., pp. 357– 411). Boston: McGraw-Hill.
Gaertner, S. L., & Dovidio, J. F. (2000). Reducing intergroup bias. The
common ingroup identity model. Philadelphia: Psychology Press.
Gonzalez-Roma, V., West, M. A., & Borrill, C. S. (2003). Demographic
diversity and team innovation: Testing a curvilinear relationship model.
Unpublished manuscript, University of Valencia.
Gruenfeld, D. H., Mannix, E. A., Williams, K. Y., & Neale, M. A. (1996).
Group composition and decision making: How member familiarity and
information distribution affect process and performance. Organizational
Behavior and Human Decision Processes, 67, 1–15.
Guzzo, R., & Dickson, M. (1996). Teams in organizations: Recent research
on performance and effectiveness. Annual Review of Psychology, 47,
307–338.
Harrison, D. A., Price, K. H., & Bell, M. P. (1998). Beyond relational
demography: Time and the effects of surface- and deep-level diversity
on work group cohesion. Academy of Management Journal, 41, 96 –107.
Harrison, D. A., Price, K. H., Gavin, J. H., & Florey, A. T. (2002). Time,
teams, and task performance: Changing effects of surface- and deep-
level diversity on group functioning. Academy of Management Journal,
45, 1029–1045.
Hewstone, M., & Brown, R. (1986). Contact is not enough: An intergroup
1020
VAN KNIPPENBERG, DE DREU, AND HOMAN

perspective on the ‘contact hypothesis.’ In M. Hewstone & R. Brown
(Eds.), Contact and conflict in intergroup encounters, (pp. 1– 44). Cam-
bridge, MA: Blackwell.
Hinsz, V. B., Tindale, R. S., & Vollrath, D. A. (1997). The emerging
conceptualization of groups as information processes. Psychological
Bulletin, 121, 43–64.
Hogg, M. A., & Abrams, D. (1988). Social identifications: A social
psychology of intergroup relations and group processes. London, United
Kingdom: Routledge.
Homan, A. C., & van Knippenberg, D. (2003, May). The beneficial effects
of cross-categorizing informational and demographical diversity in
groups. Paper presented at the 11th European Congress of Work and
Organizational Psychology, Lisbon, Portugal.
Hornsey, M. J., & Hogg, M. A. (2000). Assimilation and diversity: An
integrative model of subgroup relations. Personality and Social Psychol-
ogy Review, 4, 143–156.
Hunter, J. E., & Schmidt, F. L. (1996). Intelligence and job performance:
Economic and social implications. Psychology, Public Policy, and Law,
2, 447–472.
Jackson, S. E. (1992). Team composition in organizational settings: Issues
in managing an increasingly diverse workforce. In S. Worchel, W.
Wood, & J. A. Simpson (Eds.), Group processes and productivity (pp.
136 –180). Newbury Park, CA: Sage.
Jehn, K. (1995). A multimethod examination of the benefits and detriments
of intragroup conflict. Administrative Science Quarterly, 40, 256 –282.
Jehn, K. A., Northcraft, G. B., & Neale, M. A. (1999). Why differences
make a difference: A field study of diversity, conflict, and performance
in workgroups. Administrative Science Quarterly, 44, 741–763.
Kozlowski, S. W. J., & Bell, B. S. (2003). Work groups and teams in
organizations. In W. C. Borman & D. R. Ilgen (Eds.), Handbook of
psychology: Industrial and organizational psychology (Vol. 12, pp.
333–375). New York: Wiley.
Kruglanski, A. W., & Webster, D. M. (1991). Group members’ reactions to
opinion deviates and conformists at varying degrees of proximity to
decision deadline and of environmental noise. Journal of Personality
and Social Psychology, 61, 212–225.
Kruglanski, A. W., & Webster, D. M. (1996). Motivated closing of the
mind: “Seizing” and “freezing.” Psychological Review, 103, 263–283.
Lau, D. C., & Murnighan, J. K. (1998). Demographic diversity and fault-
lines: The compositional dynamics of organizational groups. Academy of
Management Review, 23, 325–340.
Lawrence, B. S. (1997). The black box of organizational demography.
Organization Science, 8, 1–22.
Lerner, J. S., & Tetlock, P. E. (1999). Accounting for the effects of
accountability. Psychological Bulletin, 125, 255–275.
Locke, E. A., & Latham, G. P. (1990). A theory of goal setting and task
performance. Upper Saddle River, NJ: Prentice-Hall.
Lovelace, K., Shapiro, D. L., & Weingart, L. R. (2001). Maximizing
cross-functional new product teams’ innovativeness and constraint ad-
herence: A conflict communications perspective. Academy of Manage-
ment Journal, 44, 779–793.
Mackie, D. M., Worth, L. T., & Asuncion, A. G. (1990). Processing of
persuasive in-group messages. Journal of Personality and Social Psy-
chology, 58, 812–822.
Macrae, C. N., Bodenhausen, G. V., & Milne, A. B. (1995). The dissection
of selection in person perception: Inhibitory processes in social stereo-
typing. Journal of Personality and Social Psychology, 69, 397– 407.
Marcus-Newhall, A., Miller, N., Holtz, R., & Brewer, M. B. (1993).
Cross-cutting category membership with role assignment: A means of
reducing intergroup bias. British Journal of Social Psychology, 32,
125–146.
McClelland, G. H., & Judd, C. M. (1993). Statistical difficulties of detect-
ing interactions and moderator effects. Psychological Bulletin, 114,
376 –390.
McGrath, J. E. (1984). Groups: Interaction and performance. Englewood
Cliffs, NJ: Prentice-Hall.
Milliken, F., & Martins, L. (1996). Searching for common threads: Un-
derstanding the multiple effects of diversity in organizational groups.
Academy of Management Review, 21, 402– 433.
Moreland, R. L. (1999). Transactive memory: Learning who knows what in
work groups and organizations. In L. L. Thompson, J. M. Levine, &
D. M. Messick (Eds.), Shared cognition in organizations (pp. 3–31).
Mahwah, NJ: Erlbaum.
Moss Kanter, R. (1977). Men and women of the corporation. New York:
Basic Books.
Mullen, B., & Copper, C. (1994). The relation between group cohesiveness
and performance: An integration. Psychological Bulletin, 115, 210 –227.
Murnighan, J. K., & Conlon, D. E. (1991). The dynamics of intense work
groups: A study of British string quartets. Administrative Science Quar-
terly, 36, 165–186.
Oakes, P. J., Haslam, S. A., & Turner, J. C. (1994). Stereotyping and social
reality. Malden, MA: Blackwell.
Oakes, P. J., Turner, J. C., & Haslam, S. A. (1991). Perceiving people as
group members: The role of fit in the salience of social categorizations.
British Journal of Social Psychology, 30, 125–144.
O’Reilly, C. A., Caldwell, D. F., & Barnett, W. P. (1989). Work group
demography, social integration, and turnover. Administrative Science
Quarterly, 34, 21–37.
Pelled, L. H., Eisenhardt, K. M., & Xin, K. R. (1999). Exploring the black
box: An analysis of work group diversity, conflict, and performance.
Administrative Science Quarterly, 44, 1–28.
Randel, A. E. (2002). Identity salience: A moderator of the relationship
between group gender composition and work group conflict. Journal of
Organizational Behavior, 23, 749–766.
Rhoades, L., & Eisenberger, R. (2002). Perceived organizational support:
A review of the literature. Journal of Applied Psychology, 87, 698 –714.
Riordan, C., & Shore, L. (1997). Demographic diversity and employee
attitudes: Examination of relational demography within work units.
Journal of Applied Psychology, 82, 342–358.
Schippers, M. C., Den Hartog, D. N., Koopman, P. L., & Wienk, J. A.
(2003). Reflexivity and diversity in teams: The moderating effects of
outcome interdependence and group longevity. Journal of Organiza-
tional Behavior, 24, 729–802.
Scholten, L., van Knippenberg, D., Nijstad, B. A., & De Dreu, C. K. W.
(2003). Epistemic motivation and group decision making. Unpublished
manuscript, University of Amsterdam.
Schweiger, D. M., Sandberg, W. R., & Rechner, P. L. (1989). Experiential
effects of dialectical inquiry, devil’s advocacy, and consensus ap-
proaches to strategic decision making. Academy of Management Jour-
nal, 32, 745–772.
Schwenk, C. R. (1990). Effects of devil’s advocacy and dialectical inquiry
on decision making: A meta-analysis. Organizational Behavior and
Human Decision Processes, 47, 161–176.
Simons, T., Pelled, L. H., & Smith, K. A. (1999). Making use of difference:
Diversity, debate, and decision comprehensiveness in top management
teams. Academy of Management Journal, 42, 662– 674.
Simons, T., & Peterson, R. (2000). Task conflict and relationship conflict
in top management teams: The pivotal role of intragroup trust. Journal
of Applied Psychology, 85, 102–111.
Stangor, C., Lynch, L., Duan, C., & Glass, B. (1992). Categorization of
individuals on the basis of multiple social features. Journal of Person-
ality and Social Psychology, 62, 207–218.
Stasser, G. (1999). The uncertain role of unshared information in collective
choice. In L. L. Thompson, J. M. Levine, & D. M. Messick (Eds.),
Shared cognition in organizations, (pp. 49 – 69). Mahwah, NJ: Erlbaum.
Stasser, G., Vaughan, S. I., & Stewart, D. D. (2000). Pooling unshared
information: The benefits of knowing how access to information is
distributed among group members. Organizational Behavior and Hu-
man Decision Processes, 82, 102–116.
1021
SPECIAL SECTION: WORK GROUP DIVERSITY

Stewart, G. L., & Barrick, M. R. (2000). Team structure and performance:
Assessing the mediating role of intrateam process and the moderating
role of task type. Academy of Management Journal, 43, 135–148.
Swann, W. B., Polzer, J. T., Seyle, D. C., & Ko, S. J. (2004). Finding value
in diversity: Verification of personal and social self-views in diverse
groups. Academy of Management Review, 29, 9 –27.
Tajfel, H., & Turner, J. (1986). The social identity of intergroup behavior.
In W. A. S. Worchel (Ed.), Psychology and intergroup relations. Chi-
cago: Nelson-Hall.
Thatcher, S. M. B., Jehn, K. A., & Zanutto, E. (2003). Cracks in diversity
research: The effects of diversity faultlines on conflict and performance.
Group Decision and Negotiation, 12, 217–241.
Tjosvold, D. (1998). Cooperative and competitive goal approach to con-
flict: Accomplishments and challenges. Applied Psychology: An Inter-
national Review, 47, 285–342.
Triandis, H. C., Kurowski, L. L., & Gelfand, M. J. (1994). Workplace
diversity. In H. C. Triandis, M. P. Dunnette, & L. M. Hough (Eds.),
Handbook of industrial and organizational psychology (2nd ed., Vol. 4,
pp. 769–827). Palo Alto, CA: Consulting Psychologists Press.
Tsui, A. S., Egan, T. D., & O’Reilly, C. A. (1992). Being different:
Relational demography and organizational attachment. Administrative
Science Quarterly, 37, 549–579.
Tsui, A. S., & O’Reilly, C. A. (1989). Beyond simple demographic effects:
The importance of relational demography in superior–subordinate dyads.
Academy of Management Journal, 32, 402– 423.
Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S.
(1987). Rediscovering the social group: A self-categorization theory.
Oxford, United Kingdom: Blackwell.
Tyler, T. R. (1999). Why people cooperate with organizations: An identity-
based perspective. Research in Organizational Behavior, 21, 201–246.
Van de Ven, A. H., Delbecq, A., & Koenig, R. H. (1976). Determinants of
coordination modes within organizations. American Sociological Re-
view, 41, 322–338.
van Ginkel, W., & van Knippenberg, D. (2003, September). The role of
shared mental models for informational diversity in group decision
making. Paper presented at the European Association of Experimental
Social Psychology Small Group Meeting on Small Group Decision
Making, Amsterdam.
van Knippenberg, A., & Dijksterhuis, A. (2000). Social categorization and
stereotyping: A functional perspective. European Review of Social Psy-
chology, 11, 105–144.
van Knippenberg, D. (1999). Social identity and persuasion: Reconsidering
the role of group membership. In D. Abrams & M. A. Hogg (Eds.),
Social identity and social cognition, (pp. 315–331). Oxford, United
Kingdom: Blackwell.
van Knippenberg, D. (2003). Intergroup relations in organizations. In M.
West, D. Tjosvold, & K. G. Smith (Eds.), International handbook of
organizational teamwork and cooperative working (pp. 381–399).
Chichester, United Kingdom: Wiley.
van Knippenberg, D., & Haslam, S. A. (2003). Realizing the diversity
dividend: Exploring the subtle interplay between identity, ideology, and
reality. In S. A. Haslam, D. van Knippenberg, M. J. Platow, & N.
Ellemers (Eds.), Social identity at work: Developing theory for organi-
zational practice (pp. 61–77). New York: Psychology Press.
van Knippenberg, D., Haslam, S. A., & Platow, M. J. (2003, May). Work
group diversity and work group identification: Diversity as an aspect of
group identity. Paper presented at the 11th European Congress of Work
and Organizational Psychology, Lisbon, Portugal.
van Knippenberg, D., & Hogg, M. A. (2003). A social identity model of
leadership effectiveness in organizations. Research in Organizational
Behavior, 25, 243–295.
van Knippenberg, D., & Wilke, H. (1992). Prototypicality of arguments
and conformity to in-group norms. European Journal of Social Psychol-
ogy, 22, 141–155.
van Leeuwen, E., & van Knippenberg, D. (2003). Organizational identifi-
cation following a merger: The importance of agreeing to differ. In S. A.
Haslam, D. van Knippenberg, M. J. Platow, & N. Ellemers (Eds.), Social
identity at work: Developing theory for organizational practice (pp.
205–221). New York: Psychology Press.
van Prooijen, J. W., & van Knippenberg, D. (2000). Individuation or
depersonalization: The influence of personal status position. Group
Processes and Intergroup Relations, 3, 63–77.
Van Vianen, A. E. M., & De Dreu. C. K. W. (2001). Personality in teams:
Its relationship to social cohesion, task cohesion, and team performance.
European Journal of Work and Organizational Psychology, 10, 97–120.
Wagner, W., Pfeffer, J., & O’Reilly, C. (1984). Organizational demogra-
phy and turnover in top management groups. Administrative Science
Quarterly, 29, 74–92.
Watson, W., Kumar, K., & Michaelsen, L. (1993). Cultural diversity’s
impact on interaction processes and performance: Comparing homoge-
neous and diverse task groups. Academy of Management Journal, 36,
590 – 602.
Weaver, J. L., Bowers, C. A., Salas, E., & Cannon-Bowers, J. A. (1997).
Motivation in teams. In M. M. Beyerlein & D. A. Johnson (Eds.),
Advances in interdisciplinary studies of work teams (Vol. 4, pp. 167–
191). Stamford, CT: JAI Press.
Webber, S. S., & Donahue, L. M. (2001). Impact of highly and less
job-related diversity on work group cohesion and performance: A meta-
analysis. Journal of Management, 27, 141–162.
Wegner, D. M. (1986). Transactive memory: A contemporary analysis of
the group mind. In B. Mullen & G. R. Goethals (Eds.), Theories of group
behavior (pp. 185–208). New York: Springer-Verlag.
Williams, K. Y., & O’Reilly, C. A. (1998). Demography and diversity in
organizations: A review of 40 years of research. Research in Organiza-
tional Behavior, 20, 77–140.
Wood, W. (1987). Meta-analytic review of sex differences in group per-
formance. Psychological Bulletin, 102, 53–71.
Received June 26, 2003
Revision received June 14, 2004
Accepted June 29, 2004 䡲
1022
VAN KNIPPENBERG, DE DREU, AND HOMAN