WORKPLACE DEMOGRAPHY AND ATTITUDINAL OUTCOMES: THE ROLE OF
POWER, STATUS, AND DIVERSITY CLIMATE WITHIN GROUPS
By
Abdifatah Ahmed Ali

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Psychology – Master of Arts
2015

ABSTRACT
WORKPLACE DEMOGRAPHY AND ATTITUDINAL OUTCOMES: THE ROLE OF
POWER, STATUS, AND DIVERSITY CLIMATE WITHIN GROUPS
By
Abdifatah Ahmed Ali
The study explored to what extent various environmental features can explain away the
asymmetrical effects found in the relational demography literature. We used two theoretical
paradigms stemming from different scientific disciplines: status and power dynamics from
sociology and inclusion and diversity climate from management. We posited the composition of
group members as well as those occupying managerial positions would serve as status and power
signals for out-group employees (women, minorities). These signals would in turn have
implications for the type of attitudes (i.e., job satisfaction, organizational commitment, turnover
intentions) women and minority employees endorse. Further, we posited that an inclusive climate
would inhibit social categorization processes for both in-group and out-group employees
therefore leading to better attitudinal outcomes. Across two samples, we found that the status and
power perspective did not account for the asymmetrical effects, whereas the diversity climate
perspective did in most of our hypotheses. We highlight the implications for research utilizing
both theoretical perspectives and make recommendations to companies regarding ways to
address composition and diversity climate issues within workgroups.

Copyright by
ABDIFATAH AHMED ALI
2015

ACKNOWLEDGMENTS

I would like to thank my adviser, Dr. Ann Marie Ryan, for her continual support and
guidance throughout the research process. For this, I am truly appreciative. I would also like to
thank my committee members, Drs. Daisy Chang and Neal Schmitt, for their constructive
feedback and support.
I dedicate this work to my father, Ahmed Ali Elmi, who is my main source of motivation
and inspiration.

iv

TABLE OF CONTENTS

LIST OF TABLES ...................................................................................................................... vii!
LIST OF FIGURES ................................................................................................................... viii!
INTRODUCTION ........................................................................................................................ 1!
THEORETICAL BACKGROUND ............................................................................................ 3!
Relational Demography .............................................................................................................. 3!
Conceptual Clarification in Relational Demography Literature................................................. 6!
Power, Status, and Group Dynamics ........................................................................................ 10!
Managerial Diversity and Supervisor-Subordinate Similarity.................................................. 13!
Organizational demography.................................................................................................. 13
Supervisor-subordinate similarity......................................................................................... 14!
Diversity Climate...................................................................................................................... 16!
HYPOTHESIS DEVELOPMENT ............................................................................................ 20!
The Effects of Dissimilarity...................................................................................................... 20!
Hypothesis 1.......................................................................................................................... 21!
Individual-Level Contextual Moderator ................................................................................... 21!
Supervisor-subordinate similarity......................................................................................... 21
Unit-Level Contextual Moderators........................................................................................... 22!
Managerial composition........................................................................................................ 22!
Diversity climate................................................................................................................... 24!
Asymmetry................................................................................................................................ 26!
Hypothesis 2.......................................................................................................................... 28!
Hypothesis 3…………………………………………………………………...................... 29
Research Questions................................................................................................................... 30!
Research Question 1 ............................................................................................................. 30!
Research Question 2 ............................................................................................................. 31!
SAMPLE 1: GOVERNMENT ................................................................................................... 32!
Method ...................................................................................................................................... 32!
Participants and Procedure........................................................................................................ 32!
Measures ................................................................................................................................... 33!
Diversity Climate .................................................................................................................. 33
Composition Measures.......................................................................................................... 34!
Outcome Measures................................................................................................................ 34
Controls................................................................................................................................. 35!
Analytical Strategy: Applies to Both Studies ........................................................................... 35!
Results....................................................................................................................................... 39!
Test of Hypotheses.................................................................................................................... 40!
Test of Research Questions....................................................................................................... 43!
v

SAMPLE 2: REGIONAL BANK .............................................................................................. 45!
Method ...................................................................................................................................... 45!
Participants and Procedure........................................................................................................ 45!
Measures ................................................................................................................................... 45!
Diversity Climate .................................................................................................................. 45
Composition Measures.......................................................................................................... 46!
Outcome Measures................................................................................................................ 46!
Controls................................................................................................................................. 46
Results....................................................................................................................................... 46!
Test of Hypotheses.................................................................................................................... 47!
Test of Research Questions....................................................................................................... 52!
DISCUSSION .............................................................................................................................. 54!
The Effects of Asymmetry on Attitudinal Outcomes ............................................................... 54!
Lateral comparisons. ............................................................................................................. 54
Vertical comparisons ............................................................................................................ 55!
The Effects of Diversity Climate on Attitudinal Outcomes ..................................................... 59!
Theoretical and Practical Implications ..................................................................................... 60!
Strengths, Limitations, and Future Research ............................................................................ 64!
Conclusion ................................................................................................................................ 69!
APPENDICES ............................................................................................................................. 70!
Appendix A: Tables and Figures .............................................................................................. 71!
Appendix B: Measures.............................................................................................................. 99!
REFERENCES.......................................................................................................................... 101!

vi

LIST OF TABLES
Table 1. A review of relevant articles focusing on workgroup compositions and diversity climate.
........................................................................................................................................ 71
Table 2. Means, Standard Deviations, and Correlations for Sample 1. ........................................ 75
Table 3. Hierarchical Linear Modeling Results for Job Satisfaction as the Outcome, Sample 1. 76
Table 4. Hierarchical Linear Modeling Results for Organizational Commitment as the Outcome,
Sample 1. ........................................................................................................................ 77
Table 5. Hierarchical Linear Modeling Results for Turnover Intentions as the Outcome, Sample
1. ..................................................................................................................................... 78
Table 6. Tests of Indirect Effects, Sample 1................................................................................. 79
Table 7. Means, Standard Deviations, and Correlations for Sample 2. ........................................ 80
Table 8. Hierarchical Linear Modeling Results for Job Satisfaction as the Outcome, Sample 2. 81
Table 9. Hierarchical Linear Modeling Results for Organizational Commitment as the Outcome,
Sample 2. ........................................................................................................................ 83
Table 10. Hierarchical Linear Modeling Results for Turnover Intentions as the Outcome,
Sample 2. ........................................................................................................................ 85
Table 11. Tests of Indirect Effects, Sample 2............................................................................... 87
Table 12. Summary of Hypothesis Tests Across both Samples. .................................................. 88

vii

LIST OF FIGURES

Figure 1: A pictorial representation of conceptual model. ........................................................... 89
Figure 2: Lateral versus vertical comparisons as an assessment of power and status dynamics
within groups. ................................................................................................................. 90
Figure 3: Racial diversity interacting with managerial racial diversity to predict turnover
intentions. ....................................................................................................................... 91
Figure 4. A three-way interaction between race/ethnicity, diversity climate, and racial diversity
to predict job satisfaction................................................................................................ 91
Figure 5: Diversity climate interacting with racial diversity to predict job satisfaction............... 92
Figure 6: Race/Ethnicity interacting with diversity climate to predict organizational commitment.
........................................................................................................................................ 92
Figure 7: A three-way interaction between race/ethnicity, racial diversity, and supervisorsubordinate race/ethnic similarity to predict turnover intentions. .................................. 93
Figure 8: Race/Ethnicity and supervisor-subordinate race/ethnic similarity interacting to predict
job satisfaction................................................................................................................ 93
Figure 9: Race/Ethnicity and supervisor-subordinate race/ethnic similarity interacting to predict
organizational commitment. ........................................................................................... 94
Figure 10: Gender interacting with managerial gender diversity to predict job satisfaction........ 94
Figure 11: A three-way interaction between gender, gender diversity, and diversity climate to
predict job satisfaction.................................................................................................... 95
Figure 12: A three-way interaction between gender, gender diversity, and diversity climate to
predict organizational commitment................................................................................ 95
Figure 13: A three-way interaction between gender, gender diversity, and diversity climate to
predict turnover intentions.............................................................................................. 96
Figure 14: Diversity climate and gender diversity interacting to predict job satisfaction. ........... 96
Figure 15: Diversity climate and gender diversity interacting to predict organizational
commitment.................................................................................................................... 97
Figure 16: Diversity climate and gender diversity interacting to predict turnover intentions. ..... 97
viii

Figure 17: Diversity climate and racial diversity interacting to predict turnover intentions. ....... 98

ix

INTRODUCTION
Workgroups are invariably diverse, whether that may be surface-level differences such as
educational background and age or deep-level differences, for instance values and beliefs. Such
heterogeneity in the workplace can have a profound effect on an individual’s attitudes and
behavior as well as organizational effectiveness. To this end, proponents for the business case for
diversity argue that such heterogeneity leads to beneficial outcomes (i.e., creativity and
innovation, better decision-making, customer satisfaction (Bell, Villado, Lukasik, Belau, &
Briggs, 2011; van Dijk, van Engen, & van Knippenberg, 2012). However, such proponents are
confronted with voluminous countervailing evidence highlighting the negative effects of
heterogeneity (e.g., increased conflict, performance decrements; Williams & O’Reilly, 1998) and
the tendency for individuals to engage in homophily processes (McPherson, Smith-Lovin, &
Cook, 2001). In response to the mixed findings, diversity scholars have expended a great deal of
effort into understanding the consequences of diversity within the workplace (see Hebl & Avery,
2012 for a review). A particular interest of this paper is examining the extent to which workplace
demography influences individual level outcomes (job satisfaction, organizational commitment,
and turnover intentions).
Workplace demography or “the relational and compositional demographic attributes of
individuals and collectives” (Joshi, Liao, & Roh, 2011; p. 523) has drawn attention from both
scholars and practitioners alike because of the idea that heterogeneity yields beneficial outcomes.
However, as we highlight in this paper, the accumulated knowledge in this domain provides an
unclear picture pertaining to how and when differences in demography influence outcomes. A
recent review by Joshi and colleagues urges demography scholars to connect research in

1

relational demography with sociological perspectives as a way to enhance understanding and
bring coherence to this literature.
In light of this, the aims of this study are three-fold. First, we review relevant empirical
work and theoretical perspectives that guide relational demography predictions. Second, we
introduce theoretical paradigms that are predominately in sociology and articulate their utility in
propelling our understanding of the effects of group composition. Third, we integrate the
psychological and sociological perspectives to yield specific predictions regarding how and
when group compositions should lead to various attitudinal outcomes. Finally, we address some
of the methodological criticisms in this domain by taking a multilevel perspective to studying
relational demography as well as testing the proposed hypotheses in two distinct organizational
settings.

2

THEORETICAL BACKGROUND
Relational Demography
Workplace demography and the determinants of successful group functioning have been
extensively examined in organizational science. A key theoretical perspective in this domain
deals with understanding the ways in which group characteristics affect outcomes that are
important for group success. Relational demography, defined as the degree to which individuals
are demographically similar or different from their work unit (Joshi et al., 2011) is a framework
that attempts to explain variation in attitude outcomes among demographically diverse groups.
Before the emergence of relational demography, researchers focused on attempting to explain
attitudinal differences based on nominal characteristics (e.g., gender, race/ethnicity; Mock, 1980;
Friday, Moss, & Friday, 2004). The research findings that accumulated at that point were
inconsistent, thus focusing the efforts toward group characteristics and how that might affect
outcomes across different levels of analysis (Tsui & O’Reilly, 1989).
There are three theoretical perspectives that inform relational demography research:
social identity theory, self-categorization theory, and the similarity-attraction paradigm. These
theoretical perspectives propose that individuals categorize others and themselves based on
salient characteristics (e.g., race/ethnicity, gender) and these classifications, in turn, are used to
form the basis for in-group and out-group membership (Tajfel & Turner, 1986). Further,
individuals associate positive attributes to groups they belong to and this is a source of selfesteem enhancement as well as the maintenance of positive social identity (Tajfel, 1982). One
way to sustain a positive self-view is to surround oneself with similar others as predicted by the
similarity-attraction paradigm (Byrne, 1971). As such, these theoretical frameworks underlying
relational demography make similar predictions such that higher degree of similarity on nominal

3

characteristics within a work unit yields positive outcomes. This linear and symmetrical
relationship between workgroup composition and outcomes, however, has been only partially
supported.
The next few paragraphs will highlight the empirical evidence from the relational
demography literature. The focus will be on two primary work group characteristics—gender
and race—and their relations with attitudinal outcomes. The rationale for this particular focus is
twofold: (1) the nominal characteristics mentioned have been by far the most extensive features
studied in this domain (Joshi et al., 2011; Williams & O’Reilly, 1998), and (2) attitudinal
outcomes are the primary focus of this study in that workgroup composition affects such
outcomes which then have influence on retaining employees of diverse background (Elvira &
Cohen, 2001; McKay, Avery, Tonidandel, Morris, Hernandez, & Hebl, 2007; Williams &
O’Reilly, 1998).
When the focal nominal characteristic is gender composition, research focused on
attitudinal outcomes, as in the case of job satisfaction, organizational commitment, and turnover
intentions, finds asymmetrical relationships. These converge on the idea that greater gender
similarity has positive implications for men—albeit the effect sizes are small. For instance, Tsui,
Egan, and O’Reilly, (1992) found that men reported higher organizational commitment in groups
where they were the majority. Such effects are consistent with recent findings as Gonzales and
DeNisi (2009) concluded in their study “men are less identified and committed to their
organizations the more dissimilar they are from others in their units in terms of gender, but this
did not occur for women.” (p. 30). Further, these findings remain true for job satisfaction as well
in that gender similarity is positively related to job satisfaction for men, but not for women
(Konrad, Winter, & Gutek, 1992; Smith, Smits, & Hoy, 1998; Wharton & Baron, 1991). Finally,

4

for turnover intentions, Chatman and O’Reilly (2004) found asymmetrical effects such that men
were more eager to leave work units that had higher proportions of women, but this effect did not
occur for women in male dominated work groups.
The findings for racial composition and attitudinal outcomes are less straightforward than
the conclusions drawn from research on gender composition. In certain contexts, the evidence
seems to coincide with predictions forwarded by relational demography. For instance, Zatzick,
Elvira, and Cohen (2003) found that for voluntary turnover, employees were more willing to stay
in groups that have higher proportions of similar others. Other studies, yet, find evidence that
converge on as well as diverge from relational demography predictions. Riordan and Shore
(1997) found that for African American employees, the degree of similar others they had in their
work group did affect their work group productivity as well as their perceptions of
advancements. This goes against the linearity proposition for the relationship between work
group composition and outcomes. The only consistent prediction with relational demography
was with work group commitment for African American participants. The authors provide an
explanation as to why the findings were not fully consistent with relational demography
predictions; they suggested that “it may be that affective outcomes that tap feelings of
identification and attraction, such as work group commitment, may be more likely to be affected
by relational demography for all race-ethnic groups than outcomes reflecting productivity or
advancement in the organization” (p. 354). In addition, other studies find asymmetrical effects
that show positive benefits of similarity for only White members (Hoppe, Fujishiro, & Heaney,
2013) while others such as Gonzales and DeNisi (2009) find no effect of racial composition on
attitudinal outcomes, thus complicating the true nature of the effects of racial composition.

5

Finally, the effect sizes for racial dissimilarity range from “.00 to .11 in absolute values” (Joshi
et al., 2011; p. 527).
The symmetrical effects forwarded by relational demography, which are based on social
psychology theories, may in fact be an oversimplification of how group composition influences
individual-level outcomes. Although we have gained much insight incorporating psychological
perspectives to explain intergroup dynamics, there are several issues that these theoretical
frameworks cannot fully account for regarding relational demography findings. The first issue
pertains to the prediction of linearity, such that higher similarity on nominal characteristics with
group members leads to positive outcomes, irrespective of the nature of the group under
consideration (i.e., all male versus all female employees). The empirical evidence presents a
different perspective, one that supports that more similarity is not always better—at least not for
employees who belong to low-status groups—and highlights the shortcomings of psychological
theories in addressing the full breadth of complex relations (Chattman & O’Reilly, 2004;
Riordan & Shore, 1997). The second issue considers the frequent findings of asymmetrical
effects. Indeed this issue is at the forefront of relational demography research and many scholars
have provided different explanations as to why such asymmetries exist (see Joshi et al. [2011] for
a recent discussion). A few scholars have gone so far as to suggest the asymmetrical findings to
be a methodological artifact due to unequal sample size across groups, a topic we will discuss
more in the methodology section (Tonidandel, Avery, Bucholtz, & McKay, 2008). In any case,
these limitations warrant the search for alternative explanations that could account for the
inconsistencies presented in the empirical relational demography literature.
Conceptual Clarification in Relational Demography Literature
One way to tease apart the myriad findings is to understand what diversity perspective is

6

under consideration when focusing on group composition. To do so, what is meant by diversity
has to be articulated. The typology presented by Harrison and Klein (2007) is a step in the right
direction in that it attempts to bring coherence to the diversity literature. The three types of
diversity in this typology—separation, variety, and disparity—are briefly reviewed, and their
implication for how to reconcile relational demography findings is discussed.
Separation refers to “composition of differences in (lateral) position or opinion among
unit members, primarily of value, belief, or attitude; disagreement or opposition” (Harrison &
Klein, 2007; p. 1203). Because this form of diversity is conceptualized as continuous, maximum
separation within a work-unit occurs when members are equally split at the opposite endpoint of
the continuum, as is the case when group members are split in terms of their opposing views; for
example, exactly how to distribute resources between two competing projects. Social identity
and self-categorization, and the similarity-attraction hypothesis are primary theoretical
perspectives when considering diversity as separation. As mentioned before, all these theories
forward similar predictions such that positive outcomes are expected with least separation.
Variety refers to “composition of differences in kind, source, or category of relevant
knowledge or experience among unit members; unique or distinctive information” (p. 1203).
Because of the categorical nature of variety, minimum variety occurs when all unit members
share the same category, as in the case of all members having similar functional background. On
the other end, maximum variety is achieved when unit members each represent a unique aspect
of the possible value (e.g., there is no redundancy in functional background within the work
unit). Information processing theories are important when considering variety because they
postulate diversity in information can translate to better decision-making, efficiency, and higher
productivity, among other outcomes. Thus, positive outcomes are expected with greater variety.

7

Disparity refers to “composition of (vertical) differences in proportion of socially valued
assets or resources held among unit members; inequality or relative concentration” (p. 1203).
Disparity focuses on inequality/asymmetrical distributions on a particular variable. Minimum
disparity occurs when all members have similar quantities on the focal attribute; for instance,
when all group members have uniform pay. In this example, maximum disparity would occur
when all members have low pay and one member has maximum pay. Sociological theories of
power and status are most central to the study of disparity as diversity. Thus, the extent to which
outcomes (e.g., competition versus decision-making) are positive or negative depends on the
specific form of disparity under consideration (e.g., pay disparity versus control of valued
resources).
When the focus in relational demography research is on the effects of nominal
characteristics, the conceptualization of diversity as separation appears to be the dominating
perspective (Tsui et al., 1992). Relational demography emphasizes the degree to which
individuals are different from or similar to their work-unit, with greater similarity leading to
better outcomes. As noted in the empirical evidence, however, the effects of relational
demography may not be symmetrical across nominal characteristics. Further, Harrison and Klein
(2007) make the case that when demographic attributes (e.g., gender, race, tenure, age) are the
focal consideration for examining diversity, all three conceptualizations are plausible. As such,
when different conceptualizations are meaningful, Harrison and Klein suggest “precise
specification of diversity is essential [because] it allows theorists to differentiate and compare
conceptual models, facilitating understanding and cross-fertilization and paving the way for
empirical tests of contrasting conceptions” (p. 1210).
Given that all three types of diversity seem appropriate when examining demographical

8

differences, the exact conceptual advantage or clarity, as is the case with relational demography
research, of adopting one typology over the other has to be articulated. As mentioned previously,
it appears that the separation perspective is limiting in that it does not explain why such
asymmetries exist, and therefore alternative conceptual perspectives that provide us with such
explanations are warranted. To this end, Joshi et al. (2011) discussed in their review for the need
to bridge micro and macro-level perspectives in order to enhance our understanding of relational
demography and the underlying causes for asymmetrical reactions. One way to introduce such
macro-level perspectives is to conceptualize relational demography from a disparity viewpoint.
This is useful because, as we discuss in the following section, it highlights the power and status
differentials associated with different group members. Therefore, higher degree of similarity, as
predicted by psychological theories, should have direct positive effects, in this case, attitudinal
outcomes, for in-group members (men, Whites), whereas for out-groups, the degree to which
similarity has positive effects will depend on various contextual factors that redefine power and
status relations within group members. The usage of variety as a conceptual perspective does not
apply in relational demography research because theoretical underpinnings for variety (i.e.,
resource based theory; Barney, 1991) have been used to connect with organizational-level
outcomes as is with case with firm performance and not so much on individual attitudinal
outcomes (e.g., Richard, 2000; Richard, McMillan, Chadwick, & Dwyer, 2003). The following
section illuminates the idea that relational demography research can benefit from incorporating a
disparity lens and borrowing from the sociological research on power and status. Also, Figure 1
depicts the conceptual model for this study.

9

Power, Status, and Group Dynamics
Social stratification in the form of power and status are part of organizational reality. If
we consider workgroups and their nominal (e.g., race and gender) composition to be part of a
larger system (occupation, society), it becomes difficult not to take into consideration the
distribution of status and power associated with different groups, and to the extent it has
influence on key outcomes. The effects of power and status dynamics have been the interest of
sociologists (Ridgeway, 2014; Thye, 2000), although such perspectives attract attention in
organizational science as well, but to a lesser extent (Magee & Galinsky, 2008; Ragins, 1997).
Power, in this case, is defined as “asymmetrical control over valued resources in social relations”
(Magee & Galinksy, 2008; p. 361), whereas status refers to “one’s standing in a social hierarchy
as determined by respect, deference, or social influence” (Thye, 2000; p. 408). As highlighted,
one of the primary distinctions between power and status is the focus on either control over
valued resources or the esteem value associated with individuals particularly as a function of
group membership. Historically, in the United States, minority groups—women and nonwhites—have largely been underrepresented in occupational positions that afford such power
and status (Barnett & Baron, 2000; Daley, 1996; Ridgeway, 2001a), and the differentiation along
these dimensions become meaningful when considering intergroup dynamics in the workgroup
context.
Status characteristics theory offers a rich explanation as to how nominal characteristics
become associated with differentiated levels of status and power (Ridgeway, 2001b). The theory
states that status acquisition occurs because, at the cultural level, a value premium is attached to
certain characteristics of the individual. These status characteristics lead to various levels of
performance expectation and intergroup interaction. Specifically, as to when such characteristics

10

lead to various expectations depends on whether the status is categorized as a diffuse or a specific
characteristic. Diffuse characteristics are general beliefs that “link greater social significance and
general competence, as well as specific positive and negative skills, with one category of a social
distinction compared to another” (Ridgeway, 2001b; p. 638). An instance of a diffuse
characteristic is the differential attributes ascribed to women as being more communal versus
those ascribed to men as being more agentic (Eagly & Karau, 2002). Moreover, specific status
characteristics focus on a particular domain that is relevant to the task, for example, the
consideration of auto-repairing ability when fixing a car. Thus, the theory suggests that in work
group settings, both diffuse and specific status characteristics will influence how individuals
view each other when working on collective tasks because these characteristics activate certain
expectations for performance levels. It is through these interactions that social hierarchies
emerge with varying distribution of power and status afforded to different groups.
Relational demography research utilizes social psychology frameworks to explain group
functioning. From this perspective, the categorization of who is an in-group versus an out-group
member is often times based on salient characteristics such as gender and race. Since the view of
the self is enhanced through according positive attributes to the in-group, individuals who belong
to certain categories that have positive social identities in the larger societal context (e.g., male,
White) will exert effort to maintain such identity. Indeed, social identity research does note that
the degree to which in-group members favor similar others is not true for out-group members
(Tajfel & Turner, 1986). This makes sense since higher-status employees tend to have more
power and status afforded to them in organizations than lower-status employees (Daley, 1996).
As such, it appears the expected effects of similarity on outcomes cannot be divorced from status
characteristics associated with different groups. This idea is echoed by Skaggs and DiTomaso

11

(2004) in their review of workforce diversity research when they noted, “in order to formulate
meaningful hypotheses about groups that have real status distinctions in the larger society (why
else are we interested in gender or race), it does not make sense to develop generic hypotheses
about social identity without linking these cognitive processes to the overall status structure of
the company, the industry or the occupation (e.g. “male dominated” or “female dominated”)” (p.
293). Thus, by incorporating an additional element to the relational demography paradigm—the
status characteristics of the groups—it affords us to make much more nuanced predictions about
the effects of various workgroup compositions on attitudinal outcomes.
To understand the extent to which status characteristics influence how in-group and outgroups react to workgroup composition, however, attributes of the larger organizational context
need to be included as to account for individual sensemaking activities that transpire (Skaggs &
DiTamso, 2004). As of now, the discussion has been isolated to similarity comparisons that
occur within a particular workgroup. This is to say, the focus has been on lateral comparisons—
individual attitudes are shaped by the degree of similarity with the immediate workgroup. But,
this seems to be ignoring the vertical comparison that also occur when individuals assess the
degree to which they are similar to employees at higher levels within the organization as is the
case with managers and individuals who wield power (see Figure 2 for a graphical depiction of
these various comparisons). Further, beyond similarity factors that could influence attitudinal
outcomes, we also consider the role of the environment, mainly, the extent to which diversity
climate—“shared perceptions regarding the organizational valuation of diversity and
inclusion”—has on individual outcomes (Hebl & Avery, 2012; p. 688). Accordingly, the
following section will incorporate research on organizational demography (Pfeffer, 1983) and
highlight how power and status differentials are maintained at such levels. Further, the

12

organizational diversity and inclusion literature is reviewed as it relates to relational
demography. These two contextual features in conjunction with viewing compositional diversity
from a disparity perspective are expected to provide a deeper insight about how and when
nominal characteristics within workgroups will affect attitudinal outcomes.
Managerial Diversity and Supervisor-Subordinate Similarity
Workgroups differ as a function of the various attributes associated with their nominal
characteristics. Here, it is argued that exactly when group similarity or dissimilarity leads to
better or worse outcomes depend on the demographic characteristics at higher levels within the
organization. Specifically, two vertical signals that are relevant in this situation are: (1)
organizational demography, particularly managerial composition, and (2) supervisor-subordinate
similarity. There is reason to believe such factors not only affect processes that direct the extent
to which individuals identify with their in-group as function of composition, but also factors that
contribute to the maintenance of power and status differentials.
Organizational demography. The basis of organizational demography centers around
the idea that social structures can have a profound effect on individual attitudes and behavior
(Pfeffer, 1991). Organizational actors who hold positions of power, for instance managers, affect
many outcomes (e.g., pay raises, who gets training and development) that are pertinent to
employee success (Reskin, 2000). Given that these managerial positions tend to be occupied by
the dominant in-group (e.g., male, White), it is expected through social identity mechanisms (i.e.,
status maintenance strategy), opportunities will be afforded to similar others in the organization.
Such thinking is not new and empirical evidence abounds to support that inequalities between
groups transpire as a function of in-group favoritism and homophily processes (Baldi & McBrier
1997; Reskin, McBrier, & Kmec, 1999). However, research also seems to suggest in specific

13

organizational contexts where power resides with out-group members, opportunities within the
organization tend to flow toward the enhancement of lower status employees (Cohen, Broschak,
Haveman, 1998). For example, Cohen and Broschak (2013) demonstrated using archival data
spanning 13 years that higher proportion of women managers positively related to more women
occupying created management positions. The authors connect their findings with Kanter’s
strength-in-numbers perspective that posits increases in women representation positively affects
the experience of women in the workplace—higher collective esteem—and affords them greater
accumulation of power (Kanter, 1977a). Thus, evident from the macro-level perspective and as
echoed by Joshi et al. (2011), larger organizational context and the demographic composition of
individuals in power are likely going to affect intergroup dynamics.
Supervisor-subordinate similarity. The argument for the specific focus on examining
vertical dyadic relationships stems from the idea that organizational demography does not
capture interacting agents within the workplace. Relational demographers thus have focused on
similarity between interacting agents within a particular unit (Tsui & O’Reilly, 1989). Similar to
the intergroup processes that transpire between employees who occupy the same organizational
level, the logic extends to dyadic relationships with inherent power and status asymmetries.
Thus, from this perspective, it is argued that supervisors will be attracted to subordinates who are
similar to them in nominal characteristics because it signifies in-group membership (Bryne,
1971; Tajfel & Turner, 1986), which is assumed to relate to underlying value congruence.
Because supervisors hold an authority position, it follows that in-group subordinates will benefit
in the form of higher social support, access to additional resources (e.g., enhanced organizational
network), thus positively affecting individual attitudes and behavior (Foley, Linnehan,
Greenhaus, & Weer, 2006; Jeanquart-Barone, 1996; Winfield & Rushing, 2005). These

14

similarity effects have also been found to influence performance ratings, salary, and
subordinates’ position within the organization (Avey, West, Crossley, 2008; Kraiger & Ford,
1985). For instance, in their meta-analytic review, Kraiger and Ford, (1985) found that
performance ratings were affected by supervisor-subordinate race similarity with the effects
being strong for both race groups (White and Black). Thus, it appears that such dyadic
similarities have profound effects on the experiences of subordinates, which could shape the
ways in which individuals react to group composition.
The two interrelated perspectives reviewed above suggest that organizational agents who
are in positions of power engage in behaviors that are consistent with predictions forwarded by
psychological theories. As noted by macro levels perspectives as well (see Reskin, et al. 1999),
these behaviors are one possible mechanism that reinforces power and status dynamics within
organizations. In line with macro explanations, what is important to highlight here regarding
composition is not just sheer numerical representation per se, but rather having representation in
positions that provide access to power and status. This then leads to differential distribution of
opportunities within organizations, with more afforded to the in-group, although sometimes it is
the reverse. Thus, the composition of high-status agents within organizations become rather
important because such forces are seen as one process of inequality production that could explain
the asymmetrical effects with respect to workgroup dissimilarity and attitudinal outcomes for
women and minority members (Joshi et al., 2011).
The need to examine both vertical signals is highlighted by the paucity of empirical work
attempting such integration. A literature search in top tier organizational
psychology/management (e.g., Academy of Management, Journal of Applied Psychology,
Journal of Management, Administrative Science Quarterly, Personnel Psychology, Organization

15

Science) and sociology (e.g., American Sociological Review, American Journal of Sociology)
journals since 2000 yielded 15 articles that examined either managerial diversity/composition or
supervisor-subordinate demographic similarity in its relations with various outcomes (e.g., job
performance, organizational citizenship behavior, salary, firm performance, turnover,
organizational attraction). However, none of the reviewed articles examined these vertical
signals in the same model. Further, there was only one article that focused on attitudinal
outcomes in that the researchers examined the effects of demographic similarity on satisfaction
with supervisor (Vecchio & Bullis, 2001). Examining the simultaneous effects of vertical and
lateral signals on attitudinal outcomes is one main aim of this study.
Diversity Climate
Another contextual element that is important for understanding individual behavior is
climate, which is, “the perception of formal and informal organizational policies, practices,
procedures, and routines” that occur within the organizations (Ostroff, Kinicki, & Muhammad,
2013; p. 651). Such subjective experience of the immediate environment is expected to guide
behavior and attitude formation. When these attitudes are shared amongst individual agents
within the organization, a collective phenomenon is said to emerge that could be localized within
a particular unit or can permeate throughout the organization. Various strategic or focused
climates have been studied in the organizational literature and have been linked with outcomes at
all levels in the organization. For instance, positive organizational-level service climate has been
shown to lead to increases in firm performance and higher customer service satisfaction
(Schneider, Macey, Lee, & Young, 2009). Thus, it is evident that climate is an important aspect
of the environment that has impact on organizational functioning and individual attitudes and
behavior (Schneider, Ehrhart, & Macey, 2013).

16

A specific strategic climate that could explain the asymmetrical effects of dissimilarity
within work units is diversity climate—the extent to which employees perceive their
organization values differences and institutes inclusive practices (McKay, Avery, & Morris,
2008). Here, the specific focus is on fostering an environment where individuals are valued for
their uniqueness and being different on nominal characteristics is seen beneficial to
organizational success. The literature thus far has shown ample evidence for the effects of
instilling a positive diversity climate. When organizations espouse values of inclusiveness and
establish practices in the workplace that reflect those beliefs, it is expected such situations to
improve the asymmetrical effects of dissimilarity on attitudinal outcomes.
The importance of diversity climate is highlighted by recent work providing evidence that
diversity climate “may be vital to attracting, promoting, and retaining traditionally
underrepresented employees” (Hebl & Avery, 2012; p. 689). For instance, when focusing on unit
performance, McKay and colleagues (2008, 2009) have shown that higher diversity climate leads
to smaller sales per hour mean-difference between Black-White and Hispanic-White, and when
managers and subordinates have similar pro-diversity perceptions, it tends to lead to higher sales
growth. These results provide insight into the importance of a shared climate for diversity at all
levels of the organization, and further demonstrate the monetary rewards it could potentially
yield for organizations. Extending on previous literature that reports on ethnic differences in
absenteeism, Avery, McKay, Wilson, and Tonidandel (2007) examined whether perceived
organizational value of diversity—a construct conceptually similar to diversity climate—could
attenuate the differences in absenteeism between Black-White and Hispanic-White employees.
Their results suggest when the organization values diversity, there tends to be a smaller
absenteeism difference between Blacks and Whites, but no significant difference between

17

Hispanics and Whites, thus providing some evidence for the importance of valuing diversity. The
typical effect size for diversity climate ranges from small to medium.
According to Cox’s (1994) interactional model of cultural diversity (IMCD), it is
expected that diversity climate influences individual career outcomes, as in the case of affective
(i.e., job/career satisfaction, organizational identification, and job involvement) and achievement
(i.e., job performance rating, compensation, promotion/horizontal mobility rates) outcomes,
which then influence organizational effectiveness (e.g., attendance, turnover, productivity,
etc…). Surprisingly, the vast majority of the literature on diversity climate has focused on
achievement outcomes or organizational effectiveness outcomes (see Table 1 for review). There
has not been a clear focus on affective outcomes, in particular job satisfaction, organizational
identification and commitment, and turnover intentions; studies that have examined such personcentric outcomes with respect to diversity climate have not explicitly focused on unit-level
diversity climate (Avery et al., 2007; King, Michelle, George, & Matusik, 2010; Wolfson,
Kraiger, & Finkelstein, 2011), which raises a levels-of-analysis issue. For example, of the studies
reviewed in Table 1 only 38% examined diversity climate at the unit-level, and of those studies,
none have connected diversity climate with the focal variables in this study. Even more central to
this study, the integration of the diversity climate literature in explaining asymmetrical effects
within the relational demography research has been scant. The only recognizable paper in this
area comes from Gonzalez and DeNisi (2009) who attempted to use diversity climate as a
potential buffer in examining the effects of dissimilarity on attitudinal outcomes. Their findings
suggested diversity climate attenuates the negative effects of gender dissimilarity on intentions to
quit. For nominal characteristics, African American employees reported high levels of
commitment when diversity climate was high, but these effects decreased with lower levels of

18

diversity climate. And finally, a three-way interaction highlighted that when Hispanic employees
were dissimilar from their workgroup and diversity climate was low, they reported higher levels
of turnover intentions. Although these findings provide preliminary evidence of the importance
of diversity climate in reducing dissimilarity effects, more work is needed in this area, especially
with better research designs (see Table 1 for the breakdown of study designs). Thus, in this
study, we take a multilevel perspective to studying the effects of diversity climate and not only
link it with individual-level attitudinal outcomes, but also further delineate the ways in which it
should moderate the effects of dissimilarity at the individual level.

19

HYPOTHESIS DEVELOPMENT
The Effects of Dissimilarity
Self-categorization, social identity, and similarity-attraction theories tell us we are
attracted to similar others because it enhances our self-esteem. Thus, it is expected that
individuals will respond favorably to settings where the social groups to which they belong to are
the majority. The extent to which this is the case, however, depends on the characteristics of the
social group. Specifically, whether such members belong to groups that have been traditionally
afforded higher power and status in organizations will likely influence overall identification
levels. Thus, employees who belong to dominant groups (male, White) will identify with their
group and engage in self-enhancement strategies as a way to maintain their higher social status
(Chattopadhyay, Tluchowska, & George, 2004). According to this perspective, then, we would
expect employees with dominant status to report higher positive attitudinal outcomes in work
contexts where their group is the majority. Conversely, when the focus is on members who
belong to lower status groups (e.g., women, racial minorities) the effects of similarity can be
quite complex. Social identity theory would suggest that individuals who belong to lower status
groups may engage in various identity enhancement strategies as a way to cope with the negative
attributes associated with their social groups (Tajfel & Turner, 1986). As Chattopadhyay et al.
(2004) articulated within the relational demography framework, the extent to which women and
minorities prefer workgroups that are similar or different from their social groups depends on the
particular strategy employed such that individuals who adopt a social competition and social
creativity strategy will prefer workgroups that are more homogenous, whereas those who adopt a
social mobility strategy will tend to prefer dissimilar workgroups. As an example, Chatman and
O’Reilly (2004) predicted that because of the prestige and power associated higher status groups,

20

women would prefer working in male-dominated groups. Consistent with their results, they
found that women were more likely to request a transfer from work groups that were womendominated. Further, studies focused on racial dissimilarity find either positive or negative effects
on work outcomes for minority members, thus signifying the potential use of different identity
enhancing strategies (Liao, Joshi, & Chaung, 2004; Riordan & Shore, 1997). Thus, we propose
the following generic hypothesis, which is in line with past findings. We build on this by
incorporating the contextual factors, which we posit will assist in teasing apart the exact nature
of dissimilarity effects across social groups.
Hypothesis 1. There will be an interaction between nominal characteristics (gender and
race) and workgroup composition to predict attitudinal outcomes (job satisfaction, organizational
commitment, turnover intentions). Men and White employees will be more committed, report
higher job satisfaction, and lower turnover intentions in groups that have low women and low
minority employees than women and minority employees will be in groups that have higher
women and minorities.
Individual-Level Contextual Moderator
Supervisor-subordinate similarity. Supervisors can provide a great deal of support to
their subordinates. Research findings tend to show subordinates benefit from having a supervisor
who belongs to the same social categories. For instance, gender and race congruency has been
linked with higher performance ratings (Kraiger & Ford, 1985), higher salary (Avey et al. 2008),
and receiving greater general support (Elsass & Graves, 1997) as well as specific support as is
the case with family-supportive supervision (Foley et al., 2006). These findings are consistent
with predictions advanced by social psychology perspectives in that individuals are attracted to
similar others, and in the work context, these attractions between similar others breeds in-group

21

favoritism (Brewer & Miller, 1984; Tsui & Gutek, 1999). Thus, it is expected that the
development of such connections between dyads will elicit favorable outcomes. As we have
discussed, subordinates engage in sensemaking behaviors as a way to deal with their immediate
surroundings. In a workgroup setting, the degree to which individuals react to dissimilarity may
depend on whether they share the same in-group with their supervisor. Because supervisors are
in positions of power and tend to possess more resources than the typical subordinate,
subordinates who are part of the in-group stand to gain from such a relationship. Further,
research pertaining to leader-member exchange finds that in some instances demographic
similarity leads to higher quality exchanges (Green, Anderson, & Shivers, 1996), whereas in
others, there seems to be no effect of similarity (Bauer & Green, 1996). However, in this study,
we consider supervisor-subordinate similarity as a potential moderator, in that we suggest when
individuals are dissimilar from their immediate workgroup, being similar to their supervisor
could potentially mitigate the negative effects of dissimilarity on attitudinal outcomes.
Unit-Level Contextual Moderators
Managerial composition. From a relational demography perspective, attitude formation
and behavioral influences are informed by interacting agents, as is the case with immediate
workgroup and supervisory relationships (Tsui and O’Reilly, 1989). However, the organizational
demography perspective would argue that such structural factors, for example, the demographic
composition of the organization, have direct implications on individual attitudes and behavior.
Thus a specific composition we focus on relates to managerial diversity within a particular
organizational unit. Such macro level phenomenon may prove rather useful in offering a
different lens by which we can begin to examine the differential reactions to workgroup
dissimilarity. Further, this perspective also allows for the consideration of power and status

22

dynamics within organizations, and how such forces either dismantle or maintain systems of
inequality. Similar to the argument forwarded when focusing on dyadic phenomenon, the same
goes for managerial composition in that individuals holding such positions belong to various
social groups, which could affect the functioning of the entire organization.
Diversity within the management ranks is expected to reduce the inequalities individuals
with lower status encounter in that such representation at higher levels should lead to access in a
wide array of opportunities. This perspective is connected with Kanter’s (1977a) strength-innumbers argument whereby increasing the representation of women or racial minorities reduces
inequality via a process of homosocial reproduction, which is akin to ingroup favoritism (Elliot
& Smith, 2004). For instance, the integration of women across different levels within the
organization increases when women are represented in management positions (Huffman, Cohen,
& Pearlman, 2010). In line with our argument that higher managerial diversity serves as a signal
for lower-level employees about whether certain opportunities will be granted, McGinn and
Milkman (2013) found that when there were higher proportions of senior women in the
workgroup, junior women were less likely to turnover, thus providing evidence that such signals
convey messages of mobility. Further, the authors also found a lateral comparison effect for both
women and minority junior employees such that increasing the proportion of women and
minorities in such groups led to higher turnover whereas these effects did not occur for groups
where Asian and White juniors were the majority. This finding is consistent with the idea that
access to power and status tends to decrease in settings where there is large representation of
lower status members (Mehra, Kilduff, & Brass, 1998), and as such, lower-status members will
tend to engage in distancing behaviors with their in-group (Chattopadhyay et al., 2004).
Although McGinn and Milkman’s study is informative in understanding the effects of

23

various signals (lateral versus vertical) on turnover outcomes, it is still unclear the extent to
which the outcomes might change if both signals are compared simultaneously. Along these
lines, we propose that when the proportion of lower status employees is high within a workgroup
(lateral comparison), such individuals will tend to look upward to assess the degree to which
occupants in managerial positions represent social in-groups and this will determine whether
they endorse higher withdrawal attitudes. If the vertical comparison suggests dissimilarity with
managerial composition, this will signal fewer opportunities for mobility amongst women and
minority employees, thus lowering levels of organizational commitment and job satisfaction and
increasing turnover intentions. Indeed, some preliminary findings exist to support this line of
thinking. Although Choi (2013) did not a find a main effect for managerial diversity on job
satisfaction, this relationship was moderated by race such that racial minorities reported higher
levels of satisfaction when there was higher representation of racial minorities in management
positions.
Diversity climate. Being demographically different from workgroup members can
enhance identity salience, especially when certain demographics represent a small number within
the group. Because lower status employees are considered part of the out-group, being
underrepresented could potentially lead to negative interactions with the in-group and feelings of
social exclusion (Ely, 1994; Tajfel & Turner, 1986; Tsui & O’Reilly, 1989). In such instances,
underrepresented groups may be sensitive to contextual factors that signal the organizations’
orientation toward diversity-related issues. Research in this area has documented the effects of
adopting different diversity policies could have on intergroup relations (Ely & Thomas, 2001).
The seminal work by Eli and Thomas demonstrated that when organizations institute an
integration-and-learning perspective—a perspective that is grounded on the idea that differences

24

in organizational members serves as a learning opportunity—it is effective in dealing with
intergroup conflict, therefore enhancing individual attitudes and behavior and increasing
organizational performance. The notion of diversity climate is in alignment with the integrationand-learning perspectives (Nishii, 2013), and as such believed to be significant in reducing the
negative effects associated with being dissimilar from one’s own workgroup.
Diversity management is key to intergroup relations and organizational success (Avery &
McKay, 2010). A sense of inclusion and belonging should help to reduce the effects of
dissimilarity in that inclusive environments should ease the development of a higher-order
group—the organization—in which all individuals belong (Hewstone, Martin, HammerHewstone, Crisp, & Voci, 2001). Such identification, which is engendered by an inclusive
climate, is expected to reduce conflict among different members within workgroups (Shore,
Randel, Chung, Dean, Ehrhart, & Singh, 2011). A recent study by Niishi (2013) provides
evidence for the idea that the diversity climate tends to reduce workgroup conflict as a function
of demographic diversity. Her study found that gender diversity was positively related to
relationship conflict when unit-level inclusion climate was low, while gender diversity
negatively related to relationship conflict in units where there was high inclusive climate.
Further, the study reported a positive relationship between inclusion climate and unit-level job
satisfaction and a negative relationship with unit-level turnover. These findings demonstrate the
utility in fostering an inclusive climate, in that such environments dampen the activation of status
differences that propel individuals to engage in in-group biases and out-group denigration
(Nishii, 2013; Oldmeadow, Platow, Foddy, & Anderson, 2003). In this fashion, it is expected
that the extent to which dissimilarity relates to attitudinal outcomes will be shaped by the
immediate environment, as in this case, how inclusive the workplace is.

25

Up to this point, we have forwarded predictions that help to contextualize the effects of
dissimilarity on attitudinal outcomes. A critical contribution of this work, however, is an attempt
to explain the asymmetrical effects of dissimilarity as a function of nominal characteristics in
conjunction with the contextual factors. Thus, in what follows, we propose specific moderation
effects that differ on the basis of nominal group membership.
Asymmetry
The relational demography literature is replete with asymmetrical effects across nominal
characteristics (e.g., gender, race) as a function of workgroup dissimilarity. The evidence thus
far, however, is equivocal about how and when such asymmetries emerge and particularly for
which group (Joshi et al., 2011). The contextual variables we advanced as well their theoretical
underpinnings are expected to produce useful predictions that begin to answer the questions of
how and when.
First, we propose that both managerial composition and supervisor-subordinate similarity
will produce similar effects regarding the nature of the asymmetry. To do so, we thus far have
discussed two meta-processes that are simultaneously occurring: intra and inter-group dynamics
that create and maintain status and power relations, and social group comparison (lateral versus
vertical). Organizational hierarchies are stratified based on status and power, and traditionally
higher-level positions have been accorded to White and male employees (Konrad & Gutek,
1987). In turn, it is believed that psychological and structural processes facilitate in-group
favoritism and homosocial reproduction, which not only enhances cohesion amongst in-group
members, but also further maintains the power and status asymmetries with the out-group
(Kanter, 1977b; Skaggs & DiTomaso 2004). This perspective is predominately used to explain
research findings that show positive outcomes when congruency in nominal characteristics exists

26

between supervisory relations and managerial composition (Elvira & Cohen 2001; Foley et al.,
2006; Reskin et al., 1999). Moreover, individuals engage in social comparison processes to make
sense of their environment. These social processes can take the form of comparisons in contrast
with out-group members as to assess relative standing in terms of power and status, as well as ingroup comparisons (Chattopadhyay et al. 2004; Tajfel & Turner, 1986). It is expected that such
comparisons provide useful signals regarding the extent to which individuals will engage in
various identity enhancement strategies. Given these two meta-processes, we expect
asymmetrical effects to emerge for both traditional in-group (male, White) and out-group
(women, racial minorities) members, but the nature of the asymmetry, however, will be in stark
contrast to each other between groups.
When the focus is on dominant group members, we predict that they will report higher
organizational commitment and job satisfaction, and lower turnover intentions when both lateral
and vertical differences are minimized, that is, these individuals find themselves in situations
where they are similar to their work group, have a supervisor who is part of their ingroup, and
they constitute a majority in managerial positions. Further, it is expected that there will be a
reduction in commitment and job satisfaction and increases in turnover intentions when majority
members find themselves in units where they are not well represented in positions of power and
status (supervisor and managers). When the focus is on out-group members, higher commitment
and job satisfaction and lower turnover intentions are dictated by vertical comparison as higher
out-group representation in such positions sends a signal of the possibility of upward progress
(Chattopadhyay et al. 2004; McGinn & Milkman, 2013). Conversely, women and people of color
are more likely to endorse higher turnover intentions and lower levels of organizational
commitment and job satisfaction when they are in units where they are vertically dissimilar but

27

laterally similar; such context would signal structural closure (Ibarra, 1992) and would suppress
opportunities for upward mobility, thus leading to turnover (McGinn & Milkman, 2013).
Therefore, we forward the following three-way interactions.
Hypothesis 2. There will be a three-way interaction between workgroup composition—
that is, the proportion of women and racial minorities—nominal characteristics (gender and
race), and (a) supervisor-subordinate similarity and (b) managerial composition—that is, the
proportion of women and racial minorities in managerial positions within each unit to predict
attitudinal outcomes (job satisfaction, organizational commitment, turnover intentions. The
nature of the relationship will differ based on status—that is, nominal characteristics. For men
and White members: They will report positive attitudinal outcomes when they are similar to their
unit as well higher organizational representatives and report negative attitudinal outcomes in
units where they are dissimilar to higher organizational positions. For women and racial
minorities: They will report positive attitudinal outcomes when they are dissimilar to their unit
and are similar to higher organizational representatives and report negative attitudinal outcomes
in units where they similar to their unit and are dissimilar to higher organizational
representatives.
Pertaining to diversity climate, the theoretical arguments shift from a competition and
comparison perspectives and orient the discussion around managing differences by creating a
superordinate identity—one that is shared by all members (Gaertner & Dovidio, 2000; Hewstone
et al., 2001). Organizations that are high on diversity/inclusive climate embed such ideology not
only into their formal systems (e.g., human resource practices), but also integrate it with the
social fabric of the organization (Groggins & Ryan, 2013). In such environments, organizational
identification increases because basic needs for belonging are satisfied (Shore et al., 2011). In a

28

truly inclusive climate, “both minority and majority members feel that they belong and feel
respected such that resistance and conflict are minimized” (Shore et al., 2011; p. 1277). For
instance, Hofhuis, van der Zee, and Otten (2012) found that when diversity climate was high, a
process was facilitated whereby employees from both majority and minority groups identified
with their own groups as well as their organization. Thus, environments that embrace uniqueness
as well as foster a shared identity should lead to the minimization of in-group/out-group
categorization processes that activate various status characteristics associated with different
groups. Thus, for both disadvantaged and advantaged groups, differences in attitudinal outcomes
will be pronounced in environments where diversity climate is low. Further, it is expected that
women and racial minorities will endorse lower levels of attitudinal outcomes in groups where
not only diversity climate is low, but they also constitute a larger proportion in their workgroup.
It follows that in such contexts (that is, high representation of women and racial minorities and
low diversity climate) would signal disparity, specifically the depletion of power at higher levels.
Such members might interpret this type of context as a place where there are fewer resources
afforded to diverse members and therefore fewer opportunities for upward mobility. In line with
this thinking, the following hypothesis is proposed.
Hypothesis 3. There will be a three-way interaction between workgroup composition—
that is, proportion of women and racial minorities— nominal characteristics (gender and race),
and diversity/inclusion in predicting attitudinal outcomes (job satisfaction, organizational
commitment, turnover intentions. The nature of the relationship will differ based on status—that
is, nominal characteristics. For men and White members: They will report lower attitudinal
outcomes in units where they are dissimilar from their workgroup and diversity/inclusion climate
is low. For women and racial minorities: They will report lower attitudinal outcomes in units

29

where they are similar to their workgroup and diversity/inclusion climate is low.
Research Questions
Beyond our hypothesized predictions, there are several questions we seek to answer
given our conceptual model. First, it is evident that managers, given their positional standing
within the organization, have access to resources to execute certain objectives. Research has
highlighted that managerial diversity is related to the reduction of workplace inequality
(Huffman, 2013; Huffman et al., 2010). The question then becomes: Does a diverse management
group engender a positive diversity climate? Indeed, the literature on climate in general has
established that leaders play a significant part in the development and maintenance of various
climates (Ostroff et al., 2012). Thus, it is possible that managers with diverse backgrounds use
their resources in ways that foster an inclusive climate (Avery & McKay, 2010). From an
organizational demography perspective, then, diversity climate could further serve as a mediator
explaining why having managerial diversity is positively related to attitudinal outcomes
(Lawrence, 1997). Therefore, we explore the extent to which managerial diversity is positively
related to attitudinal outcomes indirectly through the facilitation of a positive unit-level diversity
climate.
Research Question 1. Does diversity/inclusion climate mediate the relationship between
managerial composition and attitudinal outcomes?
Second, when it comes to group functioning, there is a great deal of scholarly debate as
where organizations should expend their efforts for dealing with diversity. That is to say, which
perspective—composition versus diversity climate—is more impactful in terms of individual
outcomes as well as organizational effectiveness? The conclusion drawn from empirical work
exploring both of these perspectives is equivocal (See Hebl & Avery, 2012 for a summary).

30

However, a second glance at the articles highlight possible limitations. First, most if not all the
articles examine distal outcomes that are at the unit or organizational level (Gonzales & DeNisi,
2009; McKay et al., 2008; 2009; 2012) which could suppress the true effects of composition and
climate. Second, little discussion surrounds composition and climate’s relative influence when it
comes to person-centered outcomes. Given that diversity is important because it affects
intergroup dynamics, one possible way to extend this conversation is to assess the relative
importance of unit-level attributes—workgroup and managerial composition, and diversity
climate—on attitudinal outcomes. Doing so will allow probing the question of what matters to
individuals; is it more diverse representation either at the workgroup and/or higher levels or is it
working in an environment that has a positive diversity climate or both? Thus, the following
research question will assess which predictor is more important for attitudinal outcomes using
multilevel dominance analysis (Luo & Azen, 2013).
Research Question 2. What is the relative influence of workgroup composition,
managerial composition, and diversity/inclusion climate on attitudinal outcomes?
In summary, we test our proposed model using two different samples. First, we examine
our model in the public sector setting; in this sample, we will not be testing hypothesis 2a
because such information was not gathered at the time of data collection. Second, we examine
our model in the private sector; in this sample, all aspects of the model were tested.

31

SAMPLE 1: GOVERNMENT
Method
Participants and Procedure
Participants were employees who volunteered to complete an engagement survey as part
of a larger organizational initiative during 2012. The full dataset contained a total of 20,354 nonmanagement employees across 126 units representing various agencies. The majority of
participants were White (77.2 %) and women (58.2%). All racial minority employees were
grouped together because of the sensitivity surrounding these data. All age categories were
represented: (20.6% for 34 and under), (27.8% for 35-44), (30.7% for 45-54), and (18% for 55
and over). The majority of the participants were either on the job between 3-10 years (26.6%) or
10-20 years (32.2%). Because the research questions are focused on examining the effects of
group composition on individual attitudes, units that lacked clear intact groups were deleted from
further analyses. For instance, certain units were combined to represent entire agencies, thus
making it impossible to ascertain whether employees within that unit ever interact with each
other; this is critical for computing a meaningful workgroup composition score and the
subsequent interpretation of its effects. Given these constraints, we did not include data from
participants who clearly were not part of an intact unit (i.e., units where the members composed
of it interact with each other). The reduced dataset included responses from 5,986 participants
across 101 units (average unit size = 56 employees, SD = 85; SD is skewed because of few large
units1). The race and gender proportion remained relatively the same with White and women
employees comprising 79.3% and 58.4% of the reduced sample respectively

1

We tested the extent to which the results would change as a function of deleting units with large
number of group members, defined here as units where there are more than 100 employees.
32

Measures
Diversity Climate. Five items that captured diversity climate/inclusion were used by the
organization (see Appendix A). These items overlap with those used in the diversity climate
literature (Choi, 2013; McKay et al., 2007). Using the referent shift method, individual responses
to the five diversity climate items were aggregated to the unit-level (Chan, 1998). In order to
justify aggregation, three common statistical techniques were used: within-group agreement as
calculated by the rwg and intraclass correlation (ICC) 1 and 2 (LeBreton & Senter, 2008). rwg
captures within-group agreement on the focal construct, diversity climate, with higher values
indicating stronger agreement. Aggregation is justified when rwg > .70 (LeBreton & Senter,
2008). ICC (1) provides information regarding the extent to which ratings of diversity climate
are shared by unit members. According to LeBreton and Senter (2008), a group effect of .01 is
considered small where an effect of .10 is considered medium. The authors point out that a group
effect of .05 is enough justification to aggregate to a higher level. ICC (2) captures reliability of
the group mean, and the recommended cutoff for aggregation is .70 or higher. The rwg was .68
(range = .22, .91) and the ICC (1) was .02 and ICC (2) was .53. These indices suggests that there
is moderate agreement between groups members in their ratings of diversity climate; however, a
small portion of the variance in diversity climate is at the group level and further, the groups
cannot be reliably differentiated in terms of average diversity climate (Bliese, 2000). Based on
these findings, there was not enough justification to aggregate to the group level. Therefore,
diversity climate was analyzed at the individual level. The internal consistency for this measure
was .86.

There were 15 such units. There was no difference in the pattern of relationships found across all
hypotheses.
33

Composition Measures. For group diversity, we computed the percentage of women and
minorities in each unit. Similarly, we computed the percentage of women and minorities in the
managerial team of each unit. Consistent with relational demography, workgroup composition
captures the “normative demographic standard within the group (Riordan & Wayne, 2007; p.
568) and as such, the cross-level perspective will be used. This analytical perspective has been
shown to be a superior alternative to the Euclidean distance (D-score) approach (Joshi, Liao, &
Jackson, 2006; Sacco, Scheu, Ryan, & Schmitt, 2003; Tonidandel et al. 2008). Gender diversity
is defined as the proportion of women employees in each unit. Across all 101 units, women
comprised 58 percent (range = .00 to 1; variance = .05) of work group members. Managerial
gender diversity is defined as the proportion of women (i.e., non-White) managers in each unit.
Across all 101 units, women managers comprised 46.5 percent (range = .00 to 1; variance = .06)
of management positions. Racial diversity is defined as the proportion of minority (i.e., nonWhite) employees in each unit. Across all 101 units, minorities comprised 20 percent
(range = .00 to .89; variance = .02) of work group members. Managerial racial diversity is
defined as the proportion of minority (i.e., non-White) managers in each unit. Across all 101
units, minority managers comprised 22 percent (range = .00 to 1; variance = .03) of management
positions.
Outcome Measures. For job satisfaction, three items were used (α = .66). For
organizational commitment, three items were used (α = .80), and for turnover intentions, a single
item was used (see Appendix A). The items in these measures, although not strictly coming from
specific validated measures, map onto the constructs under consideration. Such adaptations are
common in the diversity literature (Choi, 2013; Kaplan, Wiley, & Maertz, 2011; McKay et al.
2007).

34

Controls. Education was controlled because it is considered a formal status instead of a
diffuse status such as gender and race/ethnicity, which is the focus of this study (Chatman &
O’Reilly, 2004). We tested both individual education and group-level education diversity (as
measured by the Blau’s index) and only the group-level control was related to the outcomes.
Therefore, all subsequent analyses only used education diversity at the group-level.
Analytical Strategy: Applies to Both Studies
All hypotheses were examined using random coefficients modeling (RCM) with the
MPlus software package (Muthén & Muthén, 1998-2010). Below are some RCM models we
examined.
Model 1. Unconditional or Intercepts-Only Model
This model will test whether there is a non-independence issue by exploring the extent to which
variability exists within-and-between agencies/branches for the focal outcome variables. We use
turnover intentions as an exemplar.
Level-1 Model:
Turnover intentionsij = β0j + rij,
β0j = The mean turnover intentions for each agency/branch
rij = The variability in turnover intentions within each agency/branch
Level-2 Model:
β0j = γ00 + u0j,
γ00 = The grand mean for turnover intentions across all agencies/branches
u0j = The variability in turnover intentions between agencies/branches
Overall Model:
Turnover intentionsij = γ00 + u0j + rij

35

Calculating the Interclass Correlation: Test the existence of non-independence, using
the following equation.
ICC = τ2 / (τ2 + σ2) = level-2 var. / (level-2 var. + level-1 var.)
After the establishment of between agency/branch variance on the focal outcome variables, we
will proceed with testing specific hypotheses. Below are examples of testing the three-way
interactions.
Hypothesis 2a: We use turnover intentions as an outcome.
Model. Cross-Level Model
Level-1 Model:
Turnover intentionsij = β0j + β1j*(Individual Demographyij) + β2j*(SupervisorSubordinate Similarityij) + β3j* (Individual Demography* Supervisor-Subordinate
Similarityij) + rij,
Level-2 Model:
β0j = γ00 + γ01*(Group Diversityj) + u0j
β1j = γ10 + γ11*(Group Diversityj) + u1j
β2j = γ20 + γ21*(Group Diversityj) + u2j
β3j = γ30 + γ31*(Group Diversityj) + u3j
Overall Model:
Turnover Intentionsij = γ00 + γ01*Group Diversityj + γ10*Individual Demographyij +
γ11*Individual Demographyij*Group Diversityj + γ20*Supervisor-Subordinate Similarityij
+ γ21*Supervisor-Subordinate Similarityij *Group Diversityj + γ30*Individual
Demography*Supervisor-Subordinate Similarityij + γ31*Individual
Demography*Supervisor-Subordinate Similarityij*Group Diversityj + u0j +

36

u1j*Individual Demographyij + u2j*Supervisor-Subordinate Similarityij + u3j*Individual
Demography*Supervisor-Subordinate Similarityij + rij
For hypothesis 2a, we are interested in whether the cross-level, three-way interaction is
significant (γ31). Hypothesis 2b and 3 mimic the model set forth for hypothesis 2a.
The following analyses will be for the two research questions.
Test of Research Question 1: We use turnover intentions as an exemplar.
Model. 2-2-1 Multilevel Mediation
Step 1: Establish the effect of managerial diversity on turnover intentions
Level-1 Model:
Turnover intentionsij = β0j + rij,
Level-2 Model:
β0j = γ00 + γ01*(Managerial Diversityj) + u0j
Step 2: Establish the effect of managerial diversity on diversity climate
Level -2 Model:
Diversity Climatej = γ00 + γ01*(Managerial Diversityj) + u0j
Step 3: An integrated model
Level-1 Model:
Turnover intentionsij = β0j + rij,
Level-2 Model:
β0j = γ00 + γ01*(Managerial Diversityj) + γ02*(Diversity Climatej) u0j
Testing the mediation effect. A mediation effect is present when there is a significant decrease in
the coefficient associated with managerial diversity in step 1 of the mediation model as
compared to what is found for the same coefficient in the step 3 with the integrated mediation

37

model (Zhang, Zyphur, & Preacher, 2009). The Sobel (1982) test of significance will be used to
test whether the product of the two coefficients ([Step-1γ01] * [Step-3γ01]) is indeed a mediating
effect. This analysis can be executed using the Model Constraint function in Mplus.
Test of Research Question 2:
Luo and Azen (2012) forwarded a way to conduct dominance analysis within the
multilevel framework. In this methodological paper, they articulate all the various ways
predictors of interest can be assessed in terms of their relative contribution to the overall
explained variance in the outcome. Of interest to this study is the relative contribution of leveltwo variables, mainly group diversity, managerial diversity, and diversity climate in predicting
(1) the mean for the attitudinal outcomes in each unit, as well as (2) predicting individual scores
for each attitudinal outcome. These two questions will be tested using Snijder and Bosker (1994)
representation of pseudo-R2, which captures the effect size of a proposed multilevel model. To
test the first question, the

equation for level-two will be used because we are

interested in predicting the attitudinal mean score for each unit. The

equation for

predicting the mean scores in level-2 is,

where

is the level-1 variance component,

is the level-2 variance component, and n is the

unit size, and in our case because we have unequal unit sizes, we will use the harmonic mean
instead, which is computed as

, where

38

is the sample size of jth unit and j is

the number of units. The denominator in the equation pertains to the null model where there are
no predictors. As such, with this equation, we can assess the change in the
function of adding predictors. Any change in

as a

would suggest additional explanation

in the unit-level mean scores in the attitudinal outcomes (Lau & Azen, 2012). Moving forward,
in order to test our second question, the

equation for level-one will be used because

we are interesting in predicting individual attitudinal scores. The equation is as follows,

where the only difference in this equation is that we are not taking into account the average unit
size because we are not interested in predicting unit-level attitudinal means. In this case, the
will capture any change that is produced by adding an additional predictor, and thus
will be informative in that we will know how much additional variance in individual scores are
explained by adding that specific predictor. Therefore, in both of these analyses, given that we
have three predictors, we have a total of six different dominance analysis combinations (i.e., 3!).
After modeling all the different combinations, the overall average of the contributions for each
predictor will be calculated, thus giving us an idea of the general dominance of the three
variables in predicting unit-level and individual scores of attitudinal outcomes (Lau & Azen,
2012).
Results
See Table 2 for means, standard deviations, and correlations for sample 1 variables. Of
39

note, Table 2 shows that women employees reported higher organizational commitment
(r = .07, p<.01) and lower turnover intentions (r = -.04, p<.01) than men. Minority employees
reported lower diversity climate (r = -.16, p<.01), lower job satisfaction (r = -.06, p<.01), and
higher turnover intentions (r = .07, p<.01). Groups that were more gender diverse reported
higher job satisfaction (r = .04, p<.01) and organizational commitment (r = .07, p<.01), whereas
groups that were more racially diverse reported lower diversity climate (r = -.04, p<.01) but
higher organizational commitment (r = .03, p<.01). Finally, diversity climate was positively
related to job satisfaction (r = .65, p<.01) and organizational commitment (r = .64, p<.01) and
negatively related to turnover intentions (r = -.32, p<.01).
Test of Hypotheses
Before testing the hypotheses, we began with an intercept-only model for all three
outcomes: job satisfaction, organizational commitment, and turnover intentions. The ICC
coefficients were .02, .03, .01 for job satisfaction, organizational commitment, and turnover
intentions, respectively. This suggests that the majority of the variance in the attitudinal
outcomes is occurring at the within-unit level.
Hypothesis 1 stated that men and White employees would be more committed, report
higher job satisfaction, and lower turnover intentions in demographically similar group contexts
than women and racial minority group members. For gender, there was no significant interaction
between gender and gender diversity on job satisfaction (B = .15, n.s.), organizational
commitment (B = .09, n.s.), or turnover intentions (B = -.21, n.s.). For race/ethnicity, there was
also no significant interaction between race/ethnicity and racial diversity on job satisfaction
(B = -.05, n.s.), organizational commitment (B = .06, n.s.), or turnover intentions (B = .07, n.s.).
Therefore, Hypothesis 1 was not supported (see Table 3, 4, 5, for additional information when

40

the outcomes are job satisfaction, organizational commitment, and turnover intentions,
respectively).
Hypothesis 2b proposed a three-way interaction between nominal characteristics (gender
and race/ethnicity), gender and racial diversity, and managerial gender and racial diversity (for
this sample, there was no supervisor demographic information therefore we did not test
Hypothesis 2a). For gender, the three-way interaction between gender, gender diversity, and
managerial gender diversity was not significant for job satisfaction (B = -.16, n.s.), organizational
commitment (B = -.00, n.s.), or turnover intentions (B = -.26, n.s.). For race/ethnicity, the threeway interaction between race/ethnicity, racial diversity, and managerial racial diversity was not
significant for job satisfaction (B = -.34, n.s.), organizational commitment (B = -.28, n.s.), or
turnover intentions (B = -.63, n.s.). The only significant two-way interaction was between racial
diversity and managerial racial diversity in predicting turnover intentions (B = 1.36, p<.05). As
shown in Figure 3, there was a positive relationship between racial diversity and turnover when
managerial diversity was high and a negative relationship when managerial diversity was low.
Further, simple slopes analyses indicated that when managerial racial diversity was high, racial
diversity was marginally related to turnover intentions (B = .38, p = .07) but when managerial
racial diversity was low, racial diversity was not related to turnover intentions (B = -.16, n.s.).
Therefore, Hypotheses 2b was not supported (see Table 3, 4, 5 for additional information).
Hypothesis 3 proposed a three-way interaction between nominal characteristics (gender
and race/ethnicity), gender and racial diversity, and diversity climate. For gender, the three-way
interaction between gender, gender diversity, and diversity climate was not significant for job
satisfaction (B = -.05, n.s.), organizational commitment (B = -.05, n.s.), or turnover intentions
(B = .19, n.s.). For race/ethnicity, the three-way interaction between race/ethnicity, racial

41

diversity, and diversity climate was significant for job satisfaction (B = .27, p<.05). Figure 4
illustrates the nature of the interaction. Majority employees reported lower job satisfaction when
they are in units where diversity climate is low, and higher job satisfaction in units where there is
a positive diversity climate and racial diversity is low. Similar to the Majority group, minority
employees reported lower job satisfaction in units where there is a less positive diversity climate
and higher job satisfaction in units where there is a positive diversity climate and racial
composition is low. Because of the binary nature of the independent variable, we did not test for
simple slopes.
However, the three-way interaction between race/ethnicity, racial diversity, and diversity
climate was not significant for organizational commitment (B = .07, n.s.) or turnover intentions
(B = .14, n.s.). We did find, however, significant two-way interactions. First, there was a
significant interaction between diversity climate and racial diversity on job satisfaction
(B = -.48, p<.05). As shown in Figure 5, high (vs low) racial diversity matters for job satisfaction
when individuals perceive low diversity climate, but the opposite is true when individuals
perceive a positive diversity climate. Simple slopes analysis indicated that the positive
relationship between diversity climate and job satisfaction is significant at both high
(B = .75, p<.01) and low (B = .87, p<.01) levels of racial diversity. Second, there was a
significant interaction between race/ethnicity and diversity climate on organizational
commitment (B = -.06, p<.01). As shown in Figure 6, across both minority and majority
employees, organizational commitment was higher when individuals perceived higher (vs.
lower) diversity climate. However, inconsistent with prediction, majority employees reported
lower organizational commitment than minority employees when individuals perceived lower
diversity climate. Because of the binary nature of the independent variable, we did not test for

42

simple slopes. Therefore, Hypotheses 3 was partially supported (see Table 3, 4, 5 for additional
information).
Test of Research Questions
Research question one focused on testing whether managerial diversity with respect to
gender and race/ethnicity has indirect effects on the attitudinal outcomes via its effect on
diversity climate. Initially, this was conceptualized as a 2-2-1 model whereby the managerial
diversity and diversity climate were at the unit level and the outcomes were at the individual
level. However, because there was not sufficient evidence that diversity climate was a group
level variable, we changed the model to a 2-1-1 model. Essentially the model works the same as
the previous 2-2-1 model in the sense that what the mediation effect represents is a between-unit
effect. This is because at least one of our variables is at the between-unit level, in this case,
managerial diversity. Therefore, any mediation effect that is going to occur is going to happen at
the between-unit level irrespective of whether the mediating and outcome variables are at the
within-unit level (see Preacher et al. 2010 for a conceptual and mathematical explanation).
Similar to the testing of our hypotheses, we controlled for education diversity for both when
diversity climate was the outcome and when attitudinal variables were the outcomes. The
relationship between managerial diversity and diversity climate was not significant (for gender:
[B = -.03, n.s]; for race/ethnicity: [B = .06, n.s], therefore failing to establish the link necessary to
test for indirect effects. For thoroughness, we report the indirect effects for all three outcomes in
Table 6.
Research question two focused on testing a multilevel dominance analysis in order to
establish the relative importance of the three unit-level variables, group diversity, managerial
diversity, and diversity climate on the attitudinal outcomes. Because managerial and group

43

diversity did not predict the attitudinal outcomes (see Table 3, 4, 5 for outcomes job satisfaction,
organizational commitment, and turnover intentions, respectively) and diversity climate was not
analyzed at the unit level, this research question was not examined in this sample.
In sum, most of the hypotheses in our model were not supported (see Table 12 for an
overview). Next, we replicate the examination of our model using a private sector sample.

44

SAMPLE 2: REGIONAL BANK
Method
Participants and Procedure
The sample consisted of 592 customer service units of a bank that operates in
Midwestern and Eastern US states. There were a total of 6,254 participants who responded to the
2012 engagement survey. The majority of participants were White (83.4%) and female (72.5%).
Participants varied according to their organizational tenure: less than a year (17.5%); 1-2 years or
less (28%); 3-5 years (15.5%); 6-10 years (15.2%); 11-15 years (11.1%); and 16 years or more
(12.8%). The average unit size was 10.56 (SD = 48; the standard deviation is skewed because of
a few large units2).
Measures
Diversity Climate. Five items that captured diversity climate/inclusion were used by the
organization (see Appendix A). Similar to sample 1, we calculated rwg, ICC (1) and ICC (2) in
order to justify aggregating diversity climate to the group level. The rwg was .81 and the ICC (1)
was .04 and ICC (2) was .31. These indices suggests that there is strong agreement between
groups members in their ratings of diversity climate; however, a small portion of the variance in
diversity climate is at the group level and further, the groups cannot be reliably differentiated in
terms of average diversity climate. Based on these findings, there was not enough justification to
aggregate to the group level. Therefore, diversity climate was analyzed at the individual level.

2

We tested the extent to which the results would change as a function of deleting units with large
group members, defined here as units where there are more than 15 employees; there were 31
such units. The rest of the units, 561, accounted for 95 percent of the units and 47 percent of the
total sample size. There were no differences from the current findings, except when the outcome
was turnover intentions. In this particular case, the significant three-way interaction between
race/ethnicity, racial diversity, and supervisor-subordinate race similarity was not significant.
45

Internal consistency for this measure was .78.
Composition Measures. The procedure for creating group diversity and managerial
diversity was the same as for sample 1. For gender diversity, women comprised 73 percent
(range = .00 to 1; variance = .03) of work group members. For managerial gender diversity,
women managers comprised 44 percent (range = .00 to 1; variance = .08) of management
positions. For racial diversity, minorities comprised 16 percent (range = .00 to 1; variance = .03)
of work group members. For managerial racial diversity, minority managers comprised
7 percent (range = .00 to 1; variance = .02) of management positions. In addition, because this
dataset had supervisory information, we created a supervisor-subordinate similarity variable.
Specifically, we created two binary variables, one for gender and one for race/ethnicity.
Demographic similarity with supervisor was coded as 1, and dissimilarity was coded as 0. The
majority of the sample was similar to their supervisor in both gender (65 percent; SD = .47) and
race/ethnicity (77 percent; SD = .42).
Outcome Measures. For job satisfaction, three items were used (α = .83). For
organizational commitment, three items were used (α = .77), and for turnover intentions, a single
item was used (see Appendix A).
Controls. Similar to Sample 1, age and tenure, were controlled at both levels and unit
size was controlled at the unit-level. Having these controls in the model did not change the
observed pattern of relation; therefore, we present findings without them.
Results
See Table 7 for means, standard deviations, and correlations for sample 2 variables. Of
note, Table 7 shows that women employees reported lower diversity climate (r = -.06, p<.01),
higher job satisfaction (r = .03, p<.01) and lower turnover intentions (r = -.05, p<.01). Minority

46

employees reported lower diversity climate (r = -.10, p<.01), lower job satisfaction
(r = -.04, p<.01), lower organizational commitment (r = -.04, p<.01), and higher turnover
intentions (r = .06, p<.01). Women employees who were similar to their supervisor in terms of
gender reported lower organizational commitment (r = -.03, p<.01), whereas employees who
were similar to their supervisor in terms of race/ethnicity reported higher diversity climate
(r = .08, p<.01), job satisfaction (r = .05, p<.01), organizational commitment (r = .05, p<.01),
and lower turnover intentions (r = -.07, p<.01); what is important to note here is that these
relationships are mainly driven by similarity between employees who are White as they represent
a large portion of the supervisor-subordinate linkage. Groups that were more gender diverse
reported higher job satisfaction (r = .03 p<.01) and lower turnover intentions (r = -.04, p<.01),
whereas groups that were more racially diverse reported less diversity climate (r = -.06, p<.01),
less job satisfaction (r = -.04, p<.01) but lower turnover intentions (r = -.03, p<.01). Finally,
diversity climate was positively related to job satisfaction (r = .58, p<.01) and organizational
commitment (r = .65, p<.01) and negatively related to turnover intentions (r = -.36, p<.01).
Test of Hypotheses
Before testing the hypotheses, we began with an intercept-only model for all three
outcomes: job satisfaction, organizational commitment, and turnover intentions. The ICC
coefficients were .03, .05, .03 for job satisfaction, organizational commitment, and turnover
intentions, respectively. This suggests that the majority of the variance in the attitudinal
outcomes is occurring at the within-unit level.
Hypothesis 1 stated that men and White employees would be more committed, report
higher job satisfaction, and lower turnover intentions in demographically similar group contexts
than women and racial minority group members. For gender, there was no significant interaction

47

between gender and gender diversity on job satisfaction (B = -.39, n.s.), organizational
commitment (B = -.32, n.s.), or turnover intentions (B = .71, n.s.). For race/ethnicity, there was
also no significant interaction between race/ethnicity and racial diversity on job satisfaction
(B = -.33, n.s.), organizational commitment (B = .03, n.s.), or turnover intentions (B = .19, n.s.).
Therefore, Hypothesis 1 was not supported (see Table 8, 9, 10, for additional information when
the outcomes are job satisfaction, organizational commitment, and turnover intentions,
respectively).
Hypothesis 2a proposed a three-way interaction between nominal characteristics (gender
and race/ethnicity), gender and racial diversity, and supervisor-subordinate gender and
race/ethnic similarity. For gender, the three-way interaction between gender, gender diversity,
and supervisor-subordinate gender similarity was not significant for job satisfaction
(B = -.00, n.s.), organizational commitment (B = .23, n.s.), or turnover intentions
(B = -.58, n.s.). For race/ethnicity, the three-way interaction between race/ethnicity, racial
diversity, and supervisor-subordinate race/ethnic similarity was not significant for job
satisfaction (B = .66, n.s.) or organizational commitment (B = .49, n.s.). However, it was
significant for turnover intentions (B = -1.57, p<.05); Figure 7 illustrates the nature of the threeway interaction. For Majority employees, it appears their turnover intentions are in part
influenced by whether they are similar to their supervisor in race/ethnicity, irrespective of the
nature of the racial composition within their unit. However, for minority employees, they report
their highest level of turnover intentions in units where they are similar to their supervisor in
minority status and there is higher racial diversity. Further, they report lowest levels of turnover
intentions when they are dissimilar to their supervisor, irrespective of group racial diversity.

48

There were also significant two-way interactions. In particular, there was a significant interaction
between race/ethnicity and supervisor-subordinate race/ethnic similarity on job satisfaction
(B = -.18, p<.05) and organizational commitment (B = -.26, p<.01). For job satisfaction, Figure 8
illustrates the nature of the interaction. Majority employees report higher job satisfaction when
they are similar to their supervisor in race/ethnicity whereas minority employees report higher
job satisfaction when they are dissimilar to their supervisor in minority status. As can be seen in
Figure 9, the interaction effect for organizational commitment follows the same pattern as for job
satisfaction: majority employees report higher commitment when they are similar in race with
their supervisor, whereas the opposite is true for minority employees. We did not test for simple
slope effects because the variables included in this interaction were binary. Therefore, hypothesis
2a was partially supported, that is, only for the case for race/ethnicity and not for gender. Further,
the pattern of interaction for what is supported seems to be more in line with the predictions
forwarded for majority employees and not minority employees (see Table 8, 9, 10 for additional
information).
Hypothesis 2b stated that there would be a three-way interaction between nominal
characteristics (gender and race/ethnicity), gender and racial diversity, and managerial gender
and racial diversity. For gender, the three-way interaction between gender, gender diversity, and
managerial gender diversity was not significant for job satisfaction (B = .14, n.s.), organizational
commitment (B = .12, n.s.), or turnover intentions (B = .70, n.s.). For race/ethnicity, the threeway interaction between race/ethnicity, racial diversity, and managerial racial diversity was not
significant for job satisfaction (B = .88, n.s.), organizational commitment (B = -.34, n.s.), or
turnover intentions (B = .30, n.s.). However, there was a significant two-way interaction
between gender and managerial gender diversity in predicting job satisfaction (B = .16, p<.05).

49

As shown in Figure 10, men reported lower levels of job satisfaction in units where there was
higher managerial gender diversity than women. We did not test for simple slope effects because
the independent variable was binary. Thus, Hypothesis 2b was not supported (see Table 8, 9, 10,
for additional information).
Hypothesis 3 proposed a three-way interaction between nominal characteristics (gender
and race/ethnicity), gender and racial diversity, and diversity climate. For gender, the three-way
interaction between gender, gender diversity, and diversity climate was significant for job
satisfaction (B = .70, p<.05), organizational commitment (B = .62, p<.05), and turnover
intentions (B = -.91, p<.05). Figure 11 illustrates the nature of the relationship for job
satisfaction. Men reported lower job satisfaction when they are in units where diversity climate
and gender diversity are low, and higher job satisfaction in units where there is a positive
diversity climate, regardless of gender composition. Women, however, reported lower job
satisfaction in units where there is a less positive diversity climate and higher job satisfaction in
units where there is a positive diversity climate regardless of gender composition. Figure 12
illustrates the nature of the relationship for organizational commitment. Men reported lower
organizational commitment in units where diversity climate and gender diversity are low, and
higher organizational commitment in units where there was a positive diversity climate and low
gender diversity. Women reported lower organizational commitment in units where diversity
climate is low and gender diversity is high, and higher organizational commitment in units where
there is a positive diversity climate and low gender diversity. Figure 13 illustrates the nature of
the relationship for turnover intentions. Men reported higher turnover intentions when they are in
units where diversity climate and gender diversity are low, and lower turnover intentions in units
where there is a more positive diversity climate, regardless of gender composition. Women,

50

however, reported higher turnover intentions in units where there is a less positive diversity
climate and lower turnover intentions in units where there is a positive diversity climate,
regardless of gender composition. In all of these three-way interactions, we did not test for
simple slope effects because the independent variable was binary.
Further, there were also significant two-way interactions. Specifically, diversity climate
and gender diversity interacted with each other to predict job satisfaction (B = -.71, p<.01),
organizational commitment (B = -.68, p<.01), and turnover intentions (B = .98, p<.05). Figure 14
illustrates the nature of the interaction for job satisfaction. High (vs low) gender diversity matters
for job satisfaction when individuals perceive low diversity climate, but the effect of gender
diversity on job satisfaction disappears when individuals perceive a positive diversity climate.
Simple slopes analysis indicated that the positive relationship between diversity climate and job
satisfaction is significant at both high (B = .53, p<.01) and low (B = .75, p<.01) levels of gender
diversity. As can be seen in Figure 15, the interaction effect for organizational commitment is
similar in form to that of job satisfaction. Simple slopes analysis indicated that the positive
relationship between diversity climate and organizational commitment is significant at both high
(B = .59, p<.01) and low (B = .79, p<.01) levels of gender diversity. Figure 16 illustrates the
nature of the interaction for turnover intentions. Turnover intentions are at their highest when
diversity climate is low and gender diversity is low, whereas turnover intentions are at their
lowest when diversity climate is positive, regardless of the group gender diversity. Simple slopes
analysis indicated that the negative relationship between diversity climate and turnover
intentions is significant at both high (B = -.38, p<.01) and low (B = -.69, p<.01) levels of gender
diversity.

51

For race/ethnicity, the three-way interaction between race/ethnicity, racial diversity, and
diversity climate was not significant for job satisfaction (B = -.06, n.s.), organizational
commitment (B = -.13, n.s.), or turnover intentions (B = .33, n.s.). However, there was a
significant two-way interaction between racial diversity and diversity climate on turnover
intentions (B = -.51, p<.05). As can be seen in Figure 17, turnover intentions are at their highest
when there is low diversity climate and high racial diversity, whereas turnover intentions are at
their lowest when there is high racial diversity and a positive diversity climate. Simple slopes
analysis indicated that the negative relationship between diversity climate and turnover
intentions is significant at both high (B = -.61, p<.01) and low (B = -.46, p<.01) levels of racial
diversity. Therefore, Hypotheses 3 was partially supported, specifically in the case of gender and
not race/ethnicity (see Table 8, 9, 10 for additional information).
Test of Research Questions
Research question one focused on testing whether managerial diversity with respect to
gender and race/ethnicity has indirect effects on the attitudinal outcomes via its effect on
diversity climate. Similar to Sample 1, we were not able to justify the aggregation of diversity
climate to the group level, therefore we used a 2-1-1 model. The relationship between managerial
diversity and diversity climate was not significant (for gender: [B = .00, n.s]; for race/ethnicity:
[B = .07, n.s], therefore failing to establish the link necessary to test for indirect effects. For
thoroughness, we report the indirect effects for all three outcomes in Table 11.
Research question two focused on testing a multilevel dominance analysis in order to
establish the relative importance of the three unit-level variables of group diversity, managerial
diversity, and diversity climate on the attitudinal outcomes. Because managerial and group
diversity did not predict the attitudinal outcomes (see Table 8, 9, 10 for outcomes job

52

satisfaction, organizational commitment, and turnover intentions, respectively) and diversity
climate was not analyzed at the unit level, this research question was not examined.
Finally, refer to Table 12 for a summary of the results across both samples.

53

DISCUSSION
In this study, we aimed to reconcile the asymmetrical effects often found in the relational
demography literature. Specifically, we utilized a disparity lens and proposed that relational
demography cannot be fully understood without appreciating the influence of status and power
dynamics within groups. Further, we pitted the status and power perspective against the diversity
climate perspective and theorized the latter orientation would be more fruitful in accounting for
the asymmetrical effects. Below, we discuss key findings, theoretical and practical contributions,
as well as directions for future research
The Effects of Asymmetry on Attitudinal Outcomes
Lateral comparisons. We tested a baseline hypothesis that is found, albeit not
consistently, in the relational demography literature which states dissimilarity within workgroups
matters more for men and White employees than for women and minority employees. Across
both samples, this hypothesis was not supported. There are several possible interpretations for
this finding. In the case of gender, most of the asymmetrical effects found in the literature utilize
the Euclidean distance (D-score) approach to capture group dissimilarity. Tonidandel et al.
(2008) have shown such an effect could be a statistical artifact caused by range restriction in the
dissimilarity variable and may not exist in cases where the full range of distribution exists.
Further, they noted taking a multilevel approach takes into account the nesting structure of
groups, which is ignored using the D-score approach. In this study, both of these issues were
taken into consideration: our gender diversity measure displayed a normal distribution ranging
from groups that are predominately male to groups that are predominately female and we utilized
a multilevel approach. Given that we addressed the methodological concern, it is possible that
the failure to find asymmetrical effects in both of our samples may actually represent the notion

54

that men and women may not necessarily consider the gender composition of their unit as a
factor that shapes their attitudinal outcomes.
In the case of race/ethnicity, the reason for not finding asymmetrical effects is not as
clear. This is primarily due to several reasons: one, the distribution of racial diversity across units
is positively skewed. Therefore, failure to detect asymmetrical effects could have been because
of the lack of variance in our measure as similar to previous studies. Second, the grouping of all
ethnic groups into an overall minority status group is not ideal when testing for asymmetrical
effects. Such collapsing assumes any difference between groups is negligible, which is a very
strong assumption. Therefore, there is ambiguity in terms of the nature of the racial diversity
measure in that it is not a clear signal regarding one’s own ethnic group but rather a signal about
minority status employees in general. A final reason could be because of the stronger effect of
diversity climate, which is not often assessed in studies examining lateral comparisons.
Therefore, such effects might disappear when other variables such as diversity climate are
assessed within the same model (we return to this issue below when discussing the findings for
diversity climate).
Vertical comparisons. Hypothesis 2a stated that women and minority employees would
report higher attitudinal outcomes in cases where they are similar to their supervisor and there is
lower group diversity. In the case of gender, this hypothesis was not supported. We did find
partial support for this hypothesis, however, when examining race/ethnicity on turnover
intentions. As predicted, White employees reported lower intentions to leave the organizations in
contexts where they had a White supervisor. What is also important to note here is that racial
diversity did not have an effect here, that is, it did not matter whether the group composition was
more or less racially diverse, intentions to turnover were lower in all cases where White

55

employees had a White supervisor.
However, in the case of minority employees, it appears that having a supervisor who is
similar in terms of minority status actually increases intentions to leave the company. Further,
group racial diversity seems to play a role only in the case when minority employees are similar
to their supervisor in terms of minority status. Specifically, minority employee’s intentions to
leave the company are exacerbated when there is low racial diversity. This goes against upward
mobility predictions that state turnover intentions would be lower in cases where there is a strong
signal for upward movement (i.e., supervisor-subordinate similarity) and lower within-group
competition (i.e., lower group racial diversity; McGinn & Milkman, 2013). We offer a few
explanations for this finding. First, although we found evidence that is to the contrary of upward
mobility predictions, up to this point, such predictions have largely been theoretical and not
based on empirical evidence. By simultaneously considering the effects of both signals, our
findings seem to suggest racial minorities prefer units where there is less racial diversity and
their superior is White. However, this finding has to also be interpreted in light of several
limitations, which leads us to our second explanation. As noted in the results section, this data set
included units that deviated greatly from normal range. Specifically, when we analyzed the data
excluding the 31 units that were considered outliers, this three-way interaction was no longer
significant. One possible explanation for this could be the lateral signal in this case may not
reflect a workgroup signal thus creating ambiguities as to what the nature of the racial
composition actually means (i.e., does it reflect within group, within department). Moreover,
with regards to supervisor-subordinate similarity, the upward mobility signal may exist only in
cases where the subordinate is similar to the supervisor in terms of race/ethnicity. Since the
similarity we examined was minority status broadly, it is possible that it may not activate

56

identification processes connected when considering the individuals’ own ethnic group. Finally,
it could be that a minority supervisor indicates to minority employees that their function/area is
not high status, and that to achieve further status they would be better to be in a different unit.
Future research could tease apart the effect of occupying a supervisory role versus the actual
status associated with that role. On this latter point, researchers can survey incumbents regarding
their view of different supervisory positions to assess position status and power or use supervisor
salary as a proxy.
The other vertical signal we took into consideration was managerial gender and racial
diversity. The three-way hypotheses that stated nominal characteristics (gender and
race/ethnicity) would interact with group and managerial diversity were not significant.
However, there were significant effects for some of the two-way interactions. First, average-level
turnover intentions across units were in part shaped by racial group and managerial diversity.
Specifically, we found that there was a positive relationship between racial group diversity and
group-level turnover intentions in units that had higher racially diverse managers. This needs to
be interpreted in light of what the high end-points for racial group and managerial diversity mean
in each unit. These are units where managers and racial minority employees represent around 40
percent of the composition, respectively. There are two potential explanations for the emergence
of this effect at the unit level. One, in the case of majority group employees, their intentions to
leave the company would be higher in units that are diverse laterally and vertically. This
particular situation represents a case where there is a reduction in power as well as a reduction in
affiliative needs fulfillment because of the greater dissimilarity with group members. Two, in the
case of minority group members, higher representation at both levels might signal more
competition for upward mobility given that there are more minorities within each unit.

57

Therefore, it makes it difficult for differentiation, which could then increase intentions for
turnover (McGinn & Milkman, 2013).
Second, we found that gender interacted with managerial gender diversity to relate to job
satisfaction. As we would expect from social psychological theories, men reported lower job
satisfaction in units where there were more women in managerial positions. However, the
prediction set forth by sociological theories of power and status were not supported for women;
their level of job satisfaction was unaffected by the proportion of women managers. As theorized
by Chattopadhyay et al (2004), we would expect that more female representation at higher levels
within the organization would signal the possibility of upward progress and therefore, in turn,
affect the extent to which women are satisfied with their jobs. These asymmetrical effects are
found typically when considering lateral comparisons, that is, when examining the effects of
group dissimilarity. Accordingly, it is interesting to find when the comparison is vertical and has
implication for power distribution, we do not see the expected symmetry for men and women.
There may be two potential explanations. One, men may have higher expectations to be well
represented in managerial positions; therefore, in cases where this is not true, the larger deviation
from expectations may have an impact. In contrast, women may not have high expectations that
they will occupy a larger representation in managerial positions, thus their job satisfaction is
unaffected by the variability in managerial gender diversity. Second, it is possible that there is a
difference between men and women in the extent to which their current job satisfaction is
affected by upward mobility, that is, males could be looking upward when they evaluate
satisfaction with their current job (i.e., my satisfaction with what I am doing today depends on
where I think it will take me) while women are answering more based on current job (i.e., my
satisfaction ratings are based on the work I am doing right now). The idea that women’s level of

58

job satisfaction may not be as affected by upward evaluation is in part supported by work
focused on women’s managerial aspirations. A popular conception—‘Lean-In’—states women
tend not to see themselves as ready for promotion and make fewer attempts for professional
advancement and in so doing, ‘opt-out’ of pursuing leadership positions (Belkin, 2003).
However, this explanation being the principle reason for the shortages of women in leadership
positions has been challenged by a recent finding suggesting an alternative explanation that
revolves around women not receiving developmental opportunities that would position them to
be well-suited for those executive positions (Hoobler, Lemmon, Wayne, 2014). In understanding
whether this has implications for relational demography, future research should examine to what
extent asymmetrical effects such as those found here for job satisfaction are found when the
outcome is perception of promotability. We would posit that in such cases, the power signal
associated with managerial gender diversity would be stronger in affecting how women perceive
their own perceptions of upward mobility.
The Effects of Diversity Climate on Attitudinal Outcomes
Across both samples, we found that individuals with more positive perceptions of
diversity climate reported higher job satisfaction and organizational commitment and lower
turnover intentions. Furthermore, individual’s perceptions of to what extent their organization
values diversity trumps effects of group composition. For instance, we found that group
composition, in both gender and race/ethnicity mattered in situations where individuals had
lower diversity climate perceptions. In these particular situations, more women and racial
minorities in the group helped, in terms of higher job satisfaction, organizational commitment,
and lower turnover intentions. However, as perceptions of diversity climate increased, the effects
of gender and racial diversity diminished.

59

Beyond the main effect of diversity climate and its interaction with group composition,
the three-way interactions with nominal characteristics were supported. The pattern of relations
seemed similar across outcomes and across gender and race/ethnicity. The findings converged on
the idea that higher diversity climate is beneficial for both in-groups and out-groups. For
instance, we found that both men and women and majority and minority employees reported
higher job satisfaction, organizational commitment, and lower turnover intentions when they
perceived higher diversity climate, regardless of group composition. However, it is important to
note that across nominal characteristics, there was a slight trend for members to have higher
outcomes when in units that are less diverse; this can be especially seen for organizational
commitment, for example.
Theoretical and Practical Implications
This study made use of two different theoretical paradigms to help explain the
asymmetrical effects found in relational demography research. The first was the sociological
perspectives that emphasized the role of power and status associated with different group
members. To this end, we focused on diffuse status in the form of gender and race/ethnicity and
conceptualized power as the proportion of women and minorities in managerial positions and
supervisory roles. We theorized that the structural aspects of the environment would be critical in
providing information regarding the distribution of status and power within workgroups, and this
would in turn affect the extent to which out-group members react to similarity. Interestingly, as
evidenced from our findings, similar asymmetries that are typically found when considering
lateral comparisons were also found for vertical comparisons. These findings are inconsistent
with organizational demography research that finds when out-group members occupy upper
organizational levels, lower-level out-group members tend to benefit (e.g., Cohen & Broschak,

60

2013; McGinn & Milkman, 2013). However, these studies are different in that they take a
temporal aspect, therefore examining the dynamic relations between the focal variables. Our
static representation of the relationships in the current study may be obscuring the true
relationship between power distribution within groups and its effect on similarity and attitudinal
outcomes. Further, another difference between this study and those within the organizational
demography literature has to do with the criteria examined. Organizational demographers
typically examine outcomes such as promotion and voluntary turnover patterns, whereas the
focus here was on attitudinal outcomes. Thus, it is reasonable to posit that identification
processes that are targeted toward the group as well as the individual may be more important in
shaping how attitudes about job and workgroups are formed whereas the general composition of
group members may be less informative. This perspective is partially supported by the lack of
direct effect found for the composition variables on job satisfaction, organizational commitment,
and turnover intentions.
The second lens we used focused on diversity climate. Our research adds to the body of
literature in this area that highlights the benefits having a positive diversity climate (see Hebl &
Avery, 2012 for a review). While the sociological lens states that status and power differential
within groups facilitates in-group and out-group differentiation, the diversity climate perspective
is focused on the nurturing of an environment that dampens the activation of status differences
that propel individuals to engage in competition and in-group biases (Platow, Foddy, &
Anderson, 2003). Our findings lend more support to this perspective such that individuals
endorse higher levels of job satisfaction, organizational commitment, and lower levels of
turnover intentions when diversity climate is high. Furthermore, our findings suggest that group
composition does not play a large role in shaping attitudinal outcomes when diversity climate is

61

high, findings that are consistent with past literature on the importance of diversity climate (e.g.,
Gonzales & DeNisi, 2009). What is clear from our findings and in line with previous research is
when diversity climate is low, it exacerbates social categorization processes that presumably are
one possible mechanism for tension and competition within groups whereas when diversity
climate is high, it counteracts social categorization. It is apparent from the extant literature that
there are benefits of diversity climate on individual and organizational outcomes (e.g., Gonzales
& DeNisi, 2009; McKay et al., 2009; 2011; Niishi, 2013).
Finally, Niishi (2008) noted that diversity climate “facilitate[s] the engagement of whole
selves…and increasing the probability that cross-cutting ties emerge within groups” (p. 1767).
This perspective may be fruitful for team research when considering the effects of faultlines.
Such breakdowns of teams into subgroups (e.g., women versus men) are typically a function of
social categorizations processes (Lau & Murnighan, 1998). Thus, it follows that diversity climate
may play an important role in reducing the extent to which faultlines emerge among teams
because of its effects on inhibiting social categorization processes. Future research could
examine to what extent diversity climate plays a boundary condition on the emergence of
faultlines.
In light of our findings, there are several practical recommendations for organizations.
First, organizations should consider and assess the diversity climate within workgroups. The
work by Niishi (2013) suggests that diversity climate is a formative construct composed of three
dimensions: equitable employment practices, integration of differences, and inclusion in decision
making. Therefore, organizations could examine their human resource practices and ask
questions such as: how are we selecting employees/is it based on standardized metrics? Are
there biases embedded in the way we assess performance, individuals or groups? Such careful

62

consideration of human resource practices could reduce biases and increase diversity climate
perceptions. Second, continuous involvement of employees in decision-making processes could
also facilitate the development of diversity climate. Third, leadership can have an important role
in fostering a positive diversity climate. A clear pattern of relations from our findings suggests
that gender and racial diversity of managers does not necessarily affect perceptions of inclusion,
and that composition does not necessarily lead to a positive diversity climate (see Hebl & Avery
for a summary). Therefore, organizations need to think beyond not only the demographic
representation of leaders, but also what the leaders are actually doing in fostering an inclusive
atmosphere. For instance, leaders should use their social influence to support inclusion
initiatives, anticipate actions that may lead to individuals not feeling included (e.g., LMX
differentiation), and facilitate a learning environment, which is crucial for the development of a
positive diversity climate (Ely & Thomas, 2001).
Moreover, another question is how can organizations facilitate a positive diversity
climate? One possibility is to better align espoused values (e.g., value in diversity) with what is
actually in practice. Organizational members can identify whether their organization truly cares
for and values diversity versus just paying lip service to the matter. For example, Groggins and
Ryan (2013) conducted a qualitative study where they interviewed employees who worked for an
exceptionally diverse organization. A key theme that emerged from their qualitative analysis was
how deep diversity was rooted within the foundations of the organization, and because of this,
the dimensions mentioned by Niishi (2013) were easily enacted and maintained, thus becoming
an organizational reality (i.e., values in action). Therefore, organizations need to consider to what
extent diversity is central to the organizations foundation rather than being something that is
considered because of its instrumental value (i.e., increasing market share) or legal ramifications.

63

A second possibility is diversity training. A recent meta-analysis on diversity training
highlighted that diversity training has small to moderate size effects for cognitive, skill, and
affective-based outcomes (Kalinoski, Steele-Johnson, Payton, Leas, Steinke, & Bowling, 2013).
Several of their findings are critical to the current discussion. First, they found that opportunity
for social interaction during training greatly enhances affective-based outcomes. This directly
connects with the contact-hypothesis as a possible mechanism for learning and acceptance.
Second, trainee motivation played a critical role in shaping affective-based outcomes such that
trainees were very motivated about diversity training when the trainer was a direct
manager/supervisor versus other staff members, for example, diversity/inclusion manager. This
is important because it connects with the idea of legitimacy in that subordinates are more
motivated to learn and be more open to diversity issues when there is managerial support.
Therefore, diversity training could be one avenue toward making diversity a central attribute of
an organization’s identity.
Strengths, Limitations, and Future Research
Our study has several strengths that should be noted. First, we tested our model using two
samples with different settings (government versus bank) and large sample sizes at both the
within-unit as well as the between-unit level. Second, we tested and integrated theoretical
frameworks that span across levels of analyses, therefore addressing calls by Joshi et al. (2006)
to bridge micro processes (relational-demography) with macro-level influences (organizationaldemography) in order to understand asymmetrical effects. We further integrated the inclusion
literature by examining the moderating role of diversity climate. A final strength is we used a
multilevel approach to understanding the effects of asymmetry within groups and provided a

64

conservative test of our findings by simultaneously modeling our hypotheses in order to ensure
our findings were not due to chance.
Our study, however, has several limitations. In the following section, we will focus on
key limitations and ways to address them in future research.
First, the failure to find justification to aggregate the diversity climate measure might
have been a function of our items focusing on the respondent and the company and not solely at
the unit level. According to Chan (1998) this model is called the referent-shift consensus model.
The focus here is to shift the referent in the diversity climate perception from the individual to
the unit. The measurement we used across both studies did not provide a clear referent in
assessing diversity climate. For instance, some questions focused on the company (e.g., [This
company] values diversity in the workplace) whereas some questions focused on employees in
general (e.g., Employees at [This company] are able to contribute to their fullest potential
(without regard to such characteristics as age, race, ethnicity, disability, etc). Because we were
interested in unit-level diversity climate, all of the items should have had a referent to the unit
(e.g., [This unit] has a climate in which diverse perspectives are encouraged and valued).
Research that has utilized the referent-shift consensus model when examining unit-level diversity
climate has found satisfactory agreement and reliability as indexed by rwg ICC (1) and ICC (2)
(e.g., McKay et al., 2009; 2011; Niishi, 2013).
Second, work in relational demography literature has noted issues relating to the
collapsing of racial groups into a minority status category (e.g., Sacco et al., 2003; Tonidandel et
al., 2008; Vecchio & Bullis, 2001). As noted earlier, this makes a strong assumption that
different ethnic group members have a generalized ethnic identity that could be summarized
using a minority status category; essentially this suggest any difference between racially ethnic

65

minorities is insignificant and that most of the variability lies between minority and majority
status. Such generality of using White versus non-White is often, however, used for practical
purposes, as it was in our case (i.e., data files contained only such categorization) and to ease
statistical estimations. Further, in speaking to the latter issue, Tonidandel et al. (2008) noted
including additional racial categories also complicates the interpretation of findings. Finally,
another consideration and often a challenge when including multiple racial categories pertains to
having enough individuals within each group; this is important because not finding dissimilarity
effects could be due to lack of power. Further, when including additional group categories,
researchers have to ensure the unit-level group diversity variables are normally distributed, with
some units exhibiting homogeneity whereas others are more heterogeneous. This ensures that a
lack of support for asymmetry in the data cannot be attributed to a lack of variance on the focal
group composition variables (Tonidandel et al., 2008).
Third, this study is limited to the scope of the moderators examined. For instance, we
focused on illustrating the importance of vertical composition signals in helping to explain the
asymmetrical effect found in the relational demography literature. However, as evidenced across
our findings, these moderators did not help us shed light on the problem. This might be because
an implicit assumption within our framework is that ascribed status (gender, race/ethnicity) may
be informative in terms of group identification. The group identity literature, for instance, may
shed light on the complexity of group identification. Future research could examine to what
extent the asymmetrical effects could be explained by, for example, racial and gender identity.
For instance, researchers can use the multidimensional racial identity construct by Sellers, Smith,
Shelton, Rowley, and Chavous (1998), which includes racial centrality, private regard, and
public regard. The various dimensions tap into (a) locality of the race/ethnicity to one’s self-

66

concept, (b) extent to which the individuals feel good about their group, and (c) to what extent
others, that is, society holds a positive or negative view of their group. We posit that integrating
research on racial as well as gender identity with relational demography would be fruitful in
terms of illuminating under what conditions group members would be drawn toward or get away
from similar others.
Fourth, beyond moderators, our research did not propose potential explanations for the
relationship between nominal characteristics and group composition on attitudinal outcomes. As
alluded to in the previous paragraph, being similar to or different from workgroup members may
affect various identification processes that would in turn shape attitudinal outcomes. Indeed,
according to Cox’s (1994) IMCD model of diversity, organizational identification is a key
mechanism linking antecedents to individual-level outcomes. When considering relational
demography research, one might consider to what extent dissimilarity from workgroup affects
workgroup identification, which in turn affects individual outcomes such as organizational
commitment and turnover intentions. Similarly, at the unit-level, researchers can examine the
underlying processes by which group composition variables such as gender and racial diversity
affect unit-level outcomes. For instance, Niishi (2013) proposed and found support for the
importance of relationship and task conflict as primary mechanisms by which gender diversity
affects unit satisfaction and unit turnover. Finally, as noted by Hebl and Avery (2012), “it is
important to recognize that the most insightful contributions to understanding diversity’s effects
are those integrating moderation and mediation into a single model” (p. 691). Going back to
Niishi (2013) as an example, she further found climate for inclusion as an important moderator in
the process. Thus, future research should integrate both the how and the when to better guide
research on relational demography.

67

Fifth, given the cross-sectional nature of our study, it is possible that our findings could
be influenced by same-source bias (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003). However,
as noted by Siemson, Roth, and Oliveira (2010), common method variance is less of a concern
when interactions are present in the data. Also, our use of a single-item measure for turnover
intentions instead of a multiple-item measure (e.g., Cammann, Fichman, Jenkins, and Klesh’s
(1979) may have raised issues of low reliability. However, as noted by Sackett and Larson
(1990), short measures can be used when they are unambiguous in terms of what is being asked,
as is the case when asking respondents, “I intend to stay with [This company] for at least another
12 months.” On a related note, future researchers may consider using actual voluntary turnover
(e.g., Niishi, 2013) instead of intentions to quit, as the direct link between the two is contingent
on several factors (Allen, Weeks, & Moffitt, 2005).
A final limitation is the ways in which we operationalized our core variables. First,
failure to find support for our theoretical predictions regarding status and power could be a
function of the way we operationalized status and power (i.e., compositional). Had we
operationalized status in terms of perceived status, perhaps we would have found that such
perceptions are more important than the societal ascription of status. One thing to address in
future research is to what extent perceived status correlates with workgroup composition—that
is, will we find an inverse relationship between the proportion of women in the group and the
overall perceived status of the group? Second, our study did not differentiate between different
types of managerial or supervisory positions. This may be important in that different positions
provide different strength signals in terms of status and power. Therefore, future research can
code for position type and examine to what extent it influences perceptions of status and power.
Beyond the way we opertionalized status and power, we were also limited in our other, more

68

perceptual-based variables (e.g., diversity climate, attitudinal outcomes), in that our findings
could be reflective of positive affectivity. Employees who are higher on positive affect may be
more likely to view their organizational environment in a positive light, therefore influencing our
core variables. However, to the extent that our findings could be explained by such an effect is
minimal because our attitudinal variables interacted with our compositional variables to affect
attitudinal outcomes. Despite this, however, future research should include positive affectivity in
order to ensure that the findings are not a general “good place to work” effect.
Conclusion
This study answered a call made by Joshi and colleagues (2011) to integrate sociological
perspectives of status and power in order to shed light on asymmetrical effects found in
relational demography literature. Further, beyond composition, we also proposed an additional
environmental feature—diversity climate—to address our core questions. We utilized a multilevel framework and examined our hypotheses with two distinct samples. Our findings converge
on the idea that the effects of dissimilarity within workgroups across nominal characteristics are
minimized when individuals hold positive diversity climate perceptions. Thus, the diversity
climate perspective seems to better account for the asymmetrical effects than the status and
power perspective (as operationalized as composition), when considering its relations with job
satisfaction, organizational commitment, and turnover intentions.

69

APPENDICES

70

Appendix A: Tables and Figures
Table 1. A review of relevant articles focusing on workgroup compositions and diversity climate.
Authors

Acquavita et
al. (2009)

Study Focal
Variables
1.
2.
3.

Buttner et al.
(2012)

4.
5.
1.

Individual
Level
Minority status
Workplace racial
composition
Organizational
diversity efforts
Social support
Inclusion/exclusion
Diversity promise
fulfillment

Organizational
Level
None

Study Outcome
Variables
Individual
Organizational
Level
Level
1. Job
satisfaction

None

2. Turnover
intentions

None

Individual
Level
None

Organizational
Level
None

Individual
Level
86

Sample
Size
Organizational
Level
None

1. Diversity
climate (MO)
2. Organizational
Commitment
(ME)
None

None

154

None

1. Firm
motivational
cultural
intelligence
(MO)
2. Firm diversity
climate (MO)
None

305

26

175,657

191

271

26

Moderating (MO)/Mediating (ME) Variables

Chen et al.
(2012)

1. Individual
motivational
cultural intelligence

None

1. Individual
cultural sales

None

Choi (2013)

None

1. Racial/ethnic
managerial
diversity
2. Gender
managerial
diversity

1. Job satisfaction

None

1. Diversity
climate (MO)
2. Supervisory
support (MO)
3. Gender (MO)
4. Minority
(MO)

Gonzalez et
al. (2009)

1. Proportion of
gender different
others
2. Proportion of
racially/ethnically
different others
3. Gender
4. Race/Ethnicity

1. Gender
heterogeneity
2. Race/Ethnic
Heterogeneity

1. Affective
organizational
commitment
2. Organizational
identification

1. Return on
profit
2. Return on
income
3. Productivity

None

71

1.

Diversity
climate (MO)

Table 1 (cont’d).
Authors

Study Focal
Variables
Individual
Level

Herdman et al.
(2010)

None

Hofhuis et al.
(2012)

1. Diversity climate

Kossek et al.
(1993)

1. Racioethnicity
2. Gender
3. Level

Organizational
Level
1. Diversity
programs

None

1.
2.
3.
4.

Racioethnic
heterogeneity
Gender
heterogeneity
Resources for
women
Resources for
racioethnic
minorities

Study Outcome
Variables
Individual
Organizational
Level
Level
None
1. Diversity
climate

1. Job satisfaction
2. Perceived job
recognition
3. Diversityrelated conflict

None

1. Value efforts to
promote
diversity
2. Attitudes
toward
qualification of
racioethnic
minorities
3. Attitudes
toward
qualification of
women
4. Equality of
department
support of
racioethnic
minorities
5. Equality of
department
support of
women

None

72

Organizational
Level
1. Managerial
diversity (MO)
2. Managerial
relational
values (MO)
None

Individual
Level
3, 578

Sample
Size
Organizational
Level
163

1,810

None

None

775

None

Moderating (MO)/Mediating (ME) Variables
Individual
Level
None

1. Cultural
identity (ME)
2. Organizational
identity (ME)
3. Dual identity
(ME)
4. Cultural
background
(MO)
None

Table 1 (cont’d).
Authors

McKay et al.
(2007)

McKay et al.
(2008)

Study Focal
Variables

1.
2.
3.
4.
5.
1.
2.

Individual
Level
Race/Ethnicity
Gender
Tenure
Managerial Level
Diversity Climate
Race/Ethnicity
Gender

McKay et al.
(2009)

None

McKay et al.
(2011)

None

Organizational
Level
None

1. Percentage of
minority
2. Percentage of
females
3. Age diversity
1. Subordinate
diversity
climate
2. Manager
diversity
climate
3. Percent
minority
subordinates
4. Percent female
subordinates
1. Diversity
climate

Study Outcome
Variables
Individual
Organizational
Level
Level
None
1. Turnover
intentions

Individual
Level
1. Organizational
commitment
(ME)

Organizational
Level
None

Individual
Level
6,823

Sample
Size
Organizational
Level
None

Moderating (MO)/Mediating (ME) Variables

1. Sales
performance

None

None

1. Diversity
climate (MO)

6, 130

743

None

1. Sales
percentage
change (i.e.,
unit
performanc
e)-

None

1. Interactive
effects
between
subordinate
and manager
diversity
climate

56,337
subordinat
es
3,449
managers

654

None

1. Customer
satisfaction

None

1. Service
climate (MO)
2. Minority
representation
(MO)
3. Female
representation
(MO)

59,592

769

73

Table 1 (cont’d).
Authors

Study Focal
Variables
Individual
Level
1. Diversity climate

Organizational
Level
None

Mor Barak et
al. (1998)
Pugh et al.
(2008)

1. Race/Ethnicity
2. Gender
None

None

Wolfson et al.
(2011)

1. Race/Ethnicity

None

Madera et al.
(2013)

1. Workforce
racial
composition

Study Outcome
Variables
Individual
Organizational
Level
Level
1. Job satisfaction
None

Individual
Level
1. Role
ambiguity
2. Role conflict
None

Organizational
Level
None

Individual
Level
130

Sample
Size
Organizational
Level
None

None

2,686

None

1. Community
racial
composition
(MO)
None

2,369

142

1090

None

Moderating (MO)/Mediating (ME) Variables

3. Diversity
climate
None

None
1. Diversity
climate

None

1. Organizational
Commitment
2. Individual
empowerment
3. Job satisfaction

None

1. Diversity
climate (ME)

74

Table 2. Means, Standard Deviations, and Correlations for Sample 1.
Variable
1. Gender
2. Race/Ethnicity
3. Gender Diversity
4. Racial Diversity
5. Managerial
Gender Diversity
6. Managerial Racial
Diversity
7. Diversity Climate
8. Job Satisfaction
9. Organizational
Commitment
10. Turnover
Intentions

Mean

SD

1

2

3

.58

.49

--

.21
.58

.40
.22

.20

4

5

6

7

8

.10**
.45**

-.16**

--

.12

.24**

.30**

.53**

--

.47

.25

.34**

.17**

.75**

.55**

--

.22

.17

.20**

.20**

.45**

.66**

.44

--

3.62
3.39

.81
.99

-.01
.02

-.16**
-.06**

.00
.04**

-.04**
-.00

-.01
.01

3.70

.84

.07**

-.02

.07**

.03**

1.56

.83

-.04**

.07**

-.01

.01

9

-.00
.03**

-.65**

--

.04**

.05**

.64**

.70**

--

.02

.01

-.32**

-.38**

-.47**

Note. Racial diversity = the proportion of minority (i.e., non-White) employees in each unit; gender diversity = the proportion of
women employees in each unit; managerial racial diversity = (the proportion of minority (i.e., non-White) managers in each unit);
managerial gender diversity = the proportion of women managers in each unit); race/Ethnicity is coded 1= minority employees,
0 = majority employees; gender is coded 1= female employees, 0 = male employees. *p < .05; *p < .01. Scale ranged from 1–5 for
diversity climate, job satisfaction, organizational commitment, and turnover intentions.

75

Table 3. Hierarchical Linear Modeling Results for Job Satisfaction as the Outcome, Sample 1.

Job Satisfaction

Independent Variable

Intercept
Within-units Effects (Level 1)
Gender
Race/Ethnicity
Diversity Climate
Gender X Gender Diversity
Gender X Managerial Gender Diversity
Gender X Diversity Climate
Race/Ethnicity X Racial Diversity
Race/Ethnicity X Managerial Racial Diversity
Race/Ethnicity X Diversity Climate
Diversity Climate X Gender Diversity
Diversity Climate X Racial Diversity
Gender X Gender Diversity X Managerial Gender
Diversity
Race/Ethnicity X Racial Diversity X Managerial
Racial Diversity
Gender X Gender Diversity X Diversity Climate
Race/Ethnicity X Racial Diversity X Diversity
Climate
Between-Units Effects (Level 2)
Gender Diversity
Racial Diversity
Managerial Gender Diversity
Managerial Racial Diversity
Gender Diversity X Managerial Gender Diversity
Racial Diversity X Managerial Racial Diversity
Education Diversity
-2 log-likelihood
Pseudo-R-squared
Note. * p < .05. ** p < .01.

76

Model 1
β
SE
2.96**
.14

Model 2
β
SE
2.91**
.19

.03
.11**
.81**
---------

.07
.03
.02
---------

.04
.11**
.81**
.15
-.01
.01
-.05
.12
-.05
.17
-.48*

.18
.04
.03
.58
.23
.04
.37
.23
.04
.11
.20

--

--

-.16

1.52

--

--

-.34

.76

--

--

-.05

.18

--

--

.27**

.04

.05
-.19
-.11
.16
--.77**
13073.49
.41

.15
.27
.11
.15
--.25

.03
-.05
-.12
.11
.60
.31
.80**
13052.21
.41

.24
.41
.14
.31
1.17
.69
.27

Table 4. Hierarchical Linear Modeling Results for Organizational Commitment as the Outcome,
Sample 1.

Organizational Commitment

Independent Variable

Intercept
Within-units Effects (Level 1)
Gender
Race/Ethnicity
Diversity Climate
Gender X Gender Diversity
Gender X Managerial Gender Diversity
Gender X Diversity Climate
Race/Ethnicity X Racial Diversity
Race/Ethnicity X Managerial Racial Diversity
Race/Ethnicity X Diversity Climate
Diversity Climate X Gender Diversity
Diversity Climate X Racial Diversity
Gender X Gender Diversity X Managerial Gender
Diversity
Race/Ethnicity X Racial Diversity X Managerial
Racial Diversity
Gender X Gender Diversity X Diversity Climate
Race/Ethnicity X Racial Diversity X Diversity
Climate
Between-Units Effects (Level 2)
Gender Diversity
Racial Diversity
Managerial Gender Diversity
Managerial Racial Diversity
Gender Diversity X Managerial Gender Diversity
Racial Diversity X Managerial Racial Diversity
Education Diversity
-2 log-likelihood
Pseudo-R-squared
Note. * p < .05. ** p < .01.

77

Model 1
β
SE
3.37**
.10

Model 2
β
SE
3.35**
.11

.09**
.12**
.68**
---------

.02
.03
.01
---------

.09**
.11**
.70**
.09
-.07
-.01
.06
.04
-.06**
.01
-.10

.03
.04
.02
.17
.10
.03
.25
.17
.02
.10
.15

--

--

-.00

.45

--

--

-.28

.58

--

--

-.05

.14

--

--

.07

.17

.10
-.26
-.11
.08
--.56*
11194.10
.40

.24
.13
.09
.11
--.22

.08
-.27
-.08
.07
.22
.19
.52*
11184.02
.40

.20
.27
.13
.13
.51
.57
.21

Table 5. Hierarchical Linear Modeling Results for Turnover Intentions as the Outcome,
Sample 1.

Turnover Intentions

Independent Variable

Intercept
Within-units Effects (Level 1)
Gender
Race/Ethnicity
Diversity Climate
Gender X Gender Diversity
Gender X Managerial Gender Diversity
Gender X Diversity Climate
Race/Ethnicity X Racial Diversity
Race/Ethnicity X Managerial Racial Diversity
Race/Ethnicity X Diversity Climate
Diversity Climate X Gender Diversity
Diversity Climate X Racial Diversity
Gender X Gender Diversity X Managerial Gender
Diversity
Race/Ethnicity X Racial Diversity X Managerial
Racial Diversity
Gender X Gender Diversity X Diversity Climate
Race/Ethnicity X Racial Diversity X Diversity
Climate
Between-Units Effects (Level 2)
Gender Diversity
Racial Diversity
Managerial Gender Diversity
Managerial Racial Diversity
Gender Diversity X Managerial Gender Diversity
Racial Diversity X Managerial Racial Diversity
Education Diversity
-2 log-likelihood
Pseudo-R-squared
Note. * p < .05. ** p < .01.

78

Model 1
β
SE
1.56**
.07

Model 2
β
SE
1.56**
.08

-.09**
.06
-.34**
---------

.02
.05
.02
---------

-.08**
.07
-.36**
-.21
.00
.01
.07
-.26
-.00
.01
-.10

.03
.05
.03
.18
.13
.03
.25
.21
.04
.13
.16

--

--

-.16

.55

--

--

-.63

.70

--

--

.19

.14

--

--

.14

.25

-.17
.21
.19*
-.03
--.06
13569.20
.12

.09
.16
.08
.10
--.14

.01
.10
.18
-.03
.15
1.36*
.05
13552.86
.12

.13
.19
.12
.10
.47
.67
.14

Table 6. Tests of Indirect Effects, Sample 1.

Variable
Between-Person Effects
GMA ! DC ! Job Satisfaction
GMA ! DC ! Organizational Commitment
GMA ! DC ! Turnover Intentions
RMA ! DC ! Job Satisfaction
RMA ! DC ! Organizational Commitment
RMA ! DC ! Turnover Intentions

Indirect Effect
-.02
-.02
.01
.04
.03
-.02

95% Confidence Interval
Lower Bound Upper Bound
-.11
-.09
-.03
-.11
-.10
-.08

.06
.05
.04
.20
.16
.05

Note. GMA = Gender Managerial Diversity; DC = Diversity Climate; RMA = Racial Managerial
Diversity.

79

Table 7. Means, Standard Deviations, and Correlations for Sample 2.
Variable

Mean SD

1

2

3

4

5

6

7

1. Gender

.72

.44

--

2. Race/Ethnicity

.17

.37

.03**

--

.65

.47

.19**

-.02

--

.77

.42

-.01

-.68**

.03**

--

.73

.16

.36**

-.09**

.14**

.08**

--

.16

.15

-.07**

.42**

-.08**

-.33**

-.21**

--

.44

.27

.05**

-.03*

.04**

.02

.15**

-.09**

--

3. SupervisorSubordinate
Gender Similarity
4. SupervisorSubordinate Race
Similarity
5. Gender Diversity
6. Racial Diversity

8

9

10

11

7. Managerial
Gender Diversity
8. Managerial Racial
Diversity
9. Diversity Climate

.07

.13

-.04**

.11**

-.04*

-.17**

-.14**

.28**

-.05**

--

3.95

.67

-.06**

-.10**

-.02

.08**

-.01

-.06**

.01

-.00

--

10. Job Satisfaction

4.05

.73

.03*

-.04**

-.00

.05**

.03*

-.04**

.01

-.02

.58**

--

4.11

.69

-.02

-.04**

-.03*

.05**

-.02

-.02

.00

-.00

.65**

.80**

--

1.58

.89

-.05**

.06**

-.02

-.07**

-.04**

-.03**

-.00

.03*

-.36**

-.64**

-.61**

11. Organizational
Commitment
12. Turnover
Intentions

Note. Racial diversity = the proportion of minority (i.e., non-White) employees in each unit; gender diversity = the proportion of
women employees in each unit; managerial racial diversity = (the proportion of minority (i.e., non-White) managers in each unit);
managerial gender diversity = the proportion of women managers in each unit); race/Ethnicity is coded 1= minority employees,
0 = majority employees; gender is coded 1= female employees, 0 = male employees. *p < .05; *p < .01. Scale ranged from 1–5 for
diversity climate, job satisfaction, organizational commitment, and turnover intentions.

80

Table 8. Hierarchical Linear Modeling Results for Job Satisfaction as the Outcome, Sample 2.

Job Satisfaction

Independent Variable

Intercept
Within-units Effects (Level 1)
Gender
Race/Ethnicity
Diversity Climate
Supervisor-Subordinate Gender Similarity (SSGS)
Supervisor-Subordinate Race Similarity (SSRS)
Gender X Gender Diversity
Gender X Managerial Gender Diversity
Gender X Diversity Climate
Gender X SSGS
Race/Ethnicity X Racial Diversity
Race/Ethnicity X Managerial Racial Diversity
Race/Ethnicity X Diversity Climate
Race/Ethnicity X SSRS
Diversity Climate X Gender Diversity
Diversity Climate X Racial Diversity
SSGS X Gender Diversity
SSRS X Racial Diversity
Gender X Gender Diversity X Managerial Gender
Diversity
Race/Ethnicity X Racial Diversity X Managerial
Racial Diversity
Gender X Gender Diversity X Diversity Climate
Race/Ethnicity X Racial Diversity X Diversity
Climate
Gender X Gender Diversity X SSGS
Race/Ethnicity X Racial Diversity X SSRS
Note. * p < .05. ** p < .01.

81

Model 1
β
SE
3.97**
.06

Model 2
β
SE
3.99**
.05

.10**
.05
.64**
.00
.03
-------------

.01
.06
.02
.02
.04
-------------

.11**
.10
.64**
.02
.05
-.39
.16*
-.01
-.04
-.33
-.24
-.04
-.18*
-.71**
.19
-.16
-.16

.03
.07
.02
.04
.04
.35
.08
.03
.05
.43
.18
.06
.09
.26
.11
.48
.39

--

--

.14

.55

--

--

.88

.68

--

--

.70*

.30

--

--

-.06

.29

---

---

-.00
.66

.56
.54

Table 8 (cont’d).
Job Satisfaction

Independent Variable

Model 1
β
SE
Between-Units Effects (Level 2)
Gender Diversity
Racial Diversity
Managerial Gender Diversity
Managerial Racial Diversity
Gender Diversity X Managerial Gender Diversity
Racial Diversity X Managerial Racial Diversity
-2 log-likelihood
Pseudo-R-squared
Note. * p < .05. ** p < .01.

82

-.07
-.16
-.03
-.05
--8718.96
.26

.12
.12
.05
.09
---

Model 2
β
SE
.29
-.05
-.13
-.07
.02
-.04
8673.60
.27

.31
.38
.09
.09
.54
.46

Table 9. Hierarchical Linear Modeling Results for Organizational Commitment as the Outcome,
Sample 2.

Organizational Commitment

Independent Variable

Intercept
Within-units Effects (Level 1)
Gender
Race/Ethnicity
Diversity Climate
Supervisor-Subordinate Gender Similarity (SSGS)
Supervisor-Subordinate Race Similarity (SSRS)
Gender X Gender Diversity
Gender X Managerial Gender Diversity
Gender X Diversity Climate
Gender X SSGS
Race/Ethnicity X Racial Diversity
Race/Ethnicity X Managerial Racial Diversity
Race/Ethnicity X Diversity Climate
Race/Ethnicity X SSRS
Diversity Climate X Gender Diversity
Diversity Climate X Racial Diversity
SSGS X Gender Diversity
SSRS X Racial Diversity
Gender X Gender Diversity X Managerial Gender
Diversity
Race/Ethnicity X Racial Diversity X Managerial
Racial Diversity
Gender X Gender Diversity X Diversity Climate
Race/Ethnicity X Racial Diversity X Diversity
Climate
Gender X Gender Diversity X SSGS
Race/Ethnicity X Racial Diversity X SSRS
Note. * p < .05. ** p < .01.

83

Model 1
β
SE
4.11**
.04

Model 2
β
SE
4.08**
.04

.05*
.04
.69**
-.01
.02
-------------

.02
.03
.03
.03
.03
-------------

.06*
.10*
.69**
-.00
.06
-.32
.07
-.01
-.03
.03
-.07
-.01
-.26**
-.68**
15
-.25
.24

.03
.04
.03
.04
.04
.29
.08
.04
.05
.33
.14
.06
.07
.26
.11
.34
.30

--

--

.12

.48

--

--

-.34

.58

--

--

.62*

.31

--

--

-.13

.28

---

---

.23
.49

.42
.50

Table 9 (cont’d).
Organizational Commitment

Independent Variable

Model 1
β
SE
Between-Units Effects (Level 2)
Gender Diversity
Racial Diversity
Managerial Gender Diversity
Managerial Racial Diversity
Gender Diversity X Managerial Gender Diversity
Racial Diversity X Managerial Racial Diversity
-2 log-likelihood
Pseudo-R-squared
Note. * p < .05. ** p < .01.

84

-.28*
-.12
-.01
-.01
--7634.48
.35

.11
.11
.05
.08
---

Model 2
β
SE
-.02
-.32
-.05
.00
.06
.56
7592.52
.35

.23
.29
.08
.08
.46
.39

Table 10. Hierarchical Linear Modeling Results for Turnover Intentions as the Outcome,
Sample 2.

Turnover Intentions

Independent Variable

Intercept
Within-units Effects (Level 1)
Gender
Race/Ethnicity
Diversity Climate
Supervisor-Subordinate Gender Similarity (SSGS)
Supervisor-Subordinate Race Similarity (SSRS)
Gender X Gender Diversity
Gender X Managerial Gender Diversity
Gender X Diversity Climate
Gender X SSGS
Race/Ethnicity X Racial Diversity
Race/Ethnicity X Managerial Racial Diversity
Race/Ethnicity X Diversity Climate
Race/Ethnicity X SSRS
Diversity Climate X Gender Diversity
Diversity Climate X Racial Diversity
SSGS X Gender Diversity
SSRS X Racial Diversity
Gender X Gender Diversity X Managerial Gender
Diversity
Race/Ethnicity X Racial Diversity X Managerial
Racial Diversity
Gender X Gender Diversity X Diversity Climate
Race/Ethnicity X Racial Diversity X Diversity
Climate
Gender X Gender Diversity X SSGS
Race/Ethnicity X Racial Diversity X SSRS
Note. * p < .05. ** p < .01.

85

Model 1
β
SE
1.81**
.05

Model 2
β
SE
1.87**
.07

-.13**
-.01
-.54**
-.04
-.12**
-------------

.04
.05
.03
.03
.03
-------------

-.17**
-.11
-.54**
-.07
-.18**
.71
-.02
-.01
.07
.19
.29
-.02
.54
.98*
-.51*
.64
.20

.06
.08
.05
.07
.05
.46
.12
.05
.08
.49
.27
.06
.35
.40
.25
.54
.39

--

--

.70

.69

--

--

.30

1.25

--

--

-.91*

.44

--

--

.33

.45

---

---

-.58
-1.57*

.64
.72

Table 10 (cont’d).
Turnover Intentions

Independent Variable

Model 1
β
SE
Between-Units Effects (Level 2)
Gender Diversity
Racial Diversity
Managerial Gender Diversity
Managerial Racial Diversity
Gender Diversity X Managerial Gender Diversity
Racial Diversity X Managerial Racial Diversity
-2 log-likelihood
Pseudo-R-squared
Note. * p < .05. ** p < .01.

86

.00
.04
-.02
.16
--11621.32
.13

.13
.14
.06
.14
---

Model 2
β
SE
-.62
-.04
-.07
.13
-.21
-.76
11565.32
.14

.37
.40
.11
.16
.64
.92

Table 11. Tests of Indirect Effects, Sample 2.

Variable
Between-Person Effects
GMA ! DC ! Job Satisfaction
GMA ! DC ! Organizational Commitment
GMA ! DC ! Turnover Intentions
RMA ! DC ! Job Satisfaction
RMA ! DC ! Organizational Commitment
RMA ! DC ! Turnover Intentions

Indirect Effect
.00
.00
-.00
.04
.04
-.03

95% Confidence Interval
Lower Bound Upper Bound
-.05
-.06
-.04
-.04
-.04
-.09

.06
.06
.04
.12
.12
.03

Note. GMA = Gender Managerial Diversity; DC = Diversity Climate; RMA = Racial Managerial
Diversity.

87

Table 12. Summary of Hypothesis Tests Across both Samples.

Hypotheses
H1: Nominal characteristics (gender and race) will
interact with workgroup composition—that is, the
proportion of women and racial minorities—to predict
attitudinal outcomes (job satisfaction, organizational
commitment, turnover intentions). Men and White
members will more committed, report higher job
satisfaction, and lower turnover intentions in
demographically similar group context than women and
racial minority group members.
H2: There will be a three-way interaction between
workgroup composition—that is, the proportion of
women and racial minorities—nominal characteristics
(gender and race), and (a) supervisor-subordinate
similarity and (b) managerial composition—that is, the
proportion of women and racial minorities in managerial
positions within each unit to predict attitudinal outcomes
(job satisfaction, organizational commitment, turnover
intentions.
H3: There will be a three-way interaction between
workgroup composition—that is, proportion of women
and racial minorities—nominal characteristics (gender
and race), and diversity/inclusion in predicting attitudinal
outcomes (job satisfaction, organizational commitment,
turnover intentions.
Research question 1
Research question 2

Results
Sample 1:
Sample 2:
Government
Bank
Not Supported Not Supported

Not Supported

Partial support

Partial Support

Partial support

Null findings
Not tested

Null findings
Not tested

88

Notes
Sample 1:
Government

Sample 2:
Bank

Only in the case
of supervisorsubordinate
similarity and
only when the
nominal
characteristic is
race/ethnicity
Only in the case
of race/ethnicity

Only in the case
of gender

Figure 1: A pictorial representation of conceptual model.

89

Figure 2: Lateral versus vertical comparisons as an assessment of power and status dynamics
within groups.

90

Figure 3: Racial diversity interacting with managerial racial diversity to predict turnover
intentions.

Figure 4. A three-way interaction between race/ethnicity, diversity climate, and racial diversity
to predict job satisfaction.

91

Figure 5: Diversity climate interacting with racial diversity to predict job satisfaction.

Figure 6: Race/Ethnicity interacting with diversity climate to predict organizational commitment.

92

Figure 7: A three-way interaction between race/ethnicity, racial diversity, and supervisorsubordinate race/ethnic similarity to predict turnover intentions.

Figure 8: Race/Ethnicity and supervisor-subordinate race/ethnic similarity interacting to predict
job satisfaction.

93

Figure 9: Race/Ethnicity and supervisor-subordinate race/ethnic similarity interacting to predict
organizational commitment.

Figure 10: Gender interacting with managerial gender diversity to predict job satisfaction.

94

Figure 11: A three-way interaction between gender, gender diversity, and diversity climate to
predict job satisfaction.

Figure 12: A three-way interaction between gender, gender diversity, and diversity climate to
predict organizational commitment.

95

Figure 13: A three-way interaction between gender, gender diversity, and diversity climate to
predict turnover intentions.

Figure 14: Diversity climate and gender diversity interacting to predict job satisfaction.

96

Figure 15: Diversity climate and gender diversity interacting to predict organizational
commitment.

Figure 16: Diversity climate and gender diversity interacting to predict turnover intentions.

97

Figure 17: Diversity climate and racial diversity interacting to predict turnover intentions.

98

Appendix B: Measures
Government Data Measures
Diversity Climate
1. [This company] values diversity in the workplace.
2. [This company] has an inclusive work environment where individual differences are
respected.
3. My workgroup has a climate in which diverse perspectives are encouraged and valued.
4. Employees at [This company] are able to contribute to their fullest potential (without
regard to such characteristics as age, race, ethnicity, disability, etc.).
5. My colleagues treat co-workers with dignity and respect.
Job Satisfaction
1. My job gives me a feeling of personal accomplishment.
2. I believe I have the opportunity for growth in my current job.
3. I am paid fairly for the work I do.
Organizational Commitment
1. I would recommend [This company] to friends and family as a great place to work.
2. I am proud to work for the [This company].
3. My career goals can be met at [This company].
Turnover Intentions
1. I intend to stay with [This company] for at least another 12 months.
Regional Bank Data Measures
Diversity Climate
1. Everyone has a fair chance to have a successful career at [This company], regardless of
his or her differences or background.
2. At [This company], colleagues treat one another with trust and mutual respect.
3. [This company] is on the right track in creating a culture of inclusion.
4. Senior management actively supports diversity programs and events.
5. My manager works effectively with people of different kinds of backgrounds.
Job Satisfaction
1. The work I do makes me feel good about myself.
2. Considering everything, how satisfied are you with your job?
3. Considering everything, how would you rate your overall satisfaction with [This
company] at the present time?
Organizational Commitment
1. I would recommend [This company] to a friend as a place to work.
2. I feel personally committed to [This company’s] success.
3. I feel like I really belong at [This company].
99

Turnover Intentions
1. If you have your own way, will you still be working for [This company] 12 months from
now?

100

REFERENCES

101

REFERENCES
Allen, D. G., Weeks, K. P., & Moffitt, K. R. (2005). Turnover intentions and voluntary turnover:
The moderating roles of self-monitoring, locus of control, proactive personality, and risk
aversion. Journal of Applied Psychology, 90, 980-990.
Avery, D., & McKay, P. (2010). Doing diversity right: An empirically based approach to
effective diversity management. In G. Hodgkinson & J. K. Ford (Eds.), International
Review of Industrial and Organizational Psychology, (Vol. 25, pp. 227-252). West
Sussex, England: Wiley.
Avery, D. R., McKay, P. F., Wilson, D. C., & Tonidandel, S. (2007). Unequal attendance: The
relationships between race, organizational diversity cues, and absenteeism. Personnel
Psychology, 60, 875-902.
Avey, J. B., West, B. J. & Crossley, C. D. (2008). The association between ethnic congruence in
the supervisor–subordinate dyad and subordinate organizational position and salary.
Journal of Occupational and Organizational Psychology 81, 551-566.
Acquavita, S. P., Pittman, J., Gibbons, M., & Castellanos-Brown, K. (2009). Personal and
organizational diversity factors’ impact on social workers’ job satisfaction: Results from
a national Internet-based survey. Administration in Social Work, 33,151-166.
Baldi, S. & McBrier, D. B., (1997). Do the determinants of promotion differ for blacks and
whites? Evidence from the U.S. labor market. Work and Occupations 24, 478-97.
Barnett, W., Baron, J., & Stuart, T. (2000). Avenues of attainment: Occupational demography
and organizational careers in the California civil service. American Journal of Sociology,
106, 88-144.
Barney, J. (1991). Firm resources and sustained competitive advantage. Journal of
Management, 17, 99-120.
Belkin, L. 2003. The opt-out revolution. New York Times Magazine, October 26, 42-49.
Bauer, T. N., & Green, S. G. (1996). The development of leader-member exchange: A
longitudinal test. Academy of Management Journal, 39, 1538-1567.
Bell, S. T., Villado, A. J., Lukasik, M., Belau, L., & Briggs, A. (2011). Getting specific: A metaanalysis of the team demographic diversity and performance relationships. Journal of
Management, 37, 709-743.
Bliese, P (2000). Within-group agreement, non-independence, and reliability. In K. Klein & S.
Kozlowski (Eds.) Multi-level theory, research, and methods in organizations. San
Francisco: CA: Jossey-Bass (pp.349-381).
102

Buttner, E. H., Lowe, K. B., & Billing-Harris, L. (2012). An empirical test of diversity climate
dimensionality and relative effects on employee of color outcomes. Journal of Business
Ethics, 110, 247-258.
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. & Miller, N. (1984). Beyond the contact hypothesis: Theoretical perspectives on
desegregation. In N. Miller & M. B. Brewer (Eds), Groups in contact: The psychology of
desegregation, (pp. 281-302). Orlando, FL: Academic Press.
Byrne, D. (1971). The attraction paradigm. New York: Academic Press.
Cammann, C., Fichman, M., Jenkins, D., & Klesh, J. (1979). The Michigan Organizational
Assessment Questionnaire. Unpublished manuscript, University of Michigan, Ann Arbor.
Chan, D. (1998). Functional relations among constructs in the same content domain at different
levels of analysis: A typology of composition models. Journal of Applied Psychology,
83, 234-246.
Chatman, J. A., & O’Reilly, C. A. (2004). Asymmetric reactions to work group sex diversity
among men and women. Academy of management Journal, 47, 193-208.
Chattopadhyay, P., Tluchowska, M., & George, M. (2004). Identifying the ingroup: A closer
look at the influence of demographic dissimilarity on employee social identity. Academy
of Management Review, 29, 180-202.
Chen, X., Liu, D., & Portnoy, R. (2012). A multilevel investigation of motivational cultural
intelligence, organizational diversity climate, and cultural sales: evidence from U.S. real
estate firms. Journal of Applied Psychology, 97, 93-106.
Choi, S. (2013). Demographic diversity of managers and employee job satisfaction: Empirical
analysis of the federal case. Review of Public Personnel Administration, 33, 275-298.
Cohen, L. E., & Broschak, J. P. (2013). Whose jobs are these? The impact of the proportion of f
emale managers on the number of new management jobs filled by women versus men.
Administrative Science Quarterly, 58, 509-541.
Cohen, L. E., Broschak, J. P., & Haveman, H. A. (1998). And then there were more? The effects
of organizational sex composition on hiring and promotion. American Sociological
Review 63, 711-727.
Cox, T. H. Jr. (1994). Cultural diversity in organizations: Theory, research, and practice. San
Francisco: Berrett-Koehler.

103

Daley, D. (1996). Paths of glory and the glass ceiling: Differing patterns of career advancement
among women and minority federal employees. Public Administration Quarterly, 20,
143-162.
Eagly, A. H., & Karau, S. J. (2002). Role congruity theory of prejudice toward female leaders.
Psychological Review, 109, 573-598.
Elliott, J. R., & Smith, R. A. (2004). Race, gender, and workplace power. American Sociological
Review, 69, 365-386.
Elsass, P. M., & Graves, L. M. (1997). Demographic diversity in decision-making groups: The
experience of women and people of color. Academy of Management Review, 22, 946-973.
Elvira, M. M., & Cohen, L. E. (2001). Location matters: A cross-level analysis of the effects of
organizational sex composition on turnover. Academy of Management Journal, 44, 591605.
Ely, R. J., & Thomas, D. A. (2001). Cultural diversity at work: The effects of diversity
perspectives on work group processes and outcomes. Administrative Science Quarterly,
46, 229-273.
Ely, R. J. (1994). The effects of organizational demographics and social identity on relationships
among professional women. Administrative Science Quarterly 39, 203-238.
Friday, S. S., Moss, S. E., & Friday, E. (2004). Socioethnic explanations for racioethnic
differences in job satisfaction. Journal of Management Development, 23, 152-168.
Foley, S., Linnehan, F., Greenhaus, J. H., & Weer, C. H. (2006). The impact of gender similarity,
racial similarity, and work culture on family-supportive supervision. Group and
Organization Management, 31, 420-441.
Green, S. G., Anderson, S. E., & Shivers, S. L. (1996). Demographic and organizational
influence on leader–member exchange and related work attitudes. Organizational
Behavior and Human Decision Processes, 66, 203-214.
Gonzalez, J. A., & DeNisi, A. S. (2009). Cross-level effects of demography and diversity climate
on organizational attachment and firm effectiveness. Journal of Organizational Behavior,
30, 21-40.
Groggins, A., & Ryan, A. M. (2013). Embracing uniqueness: The underpinnings of a positive
climate for diversity. Journal of Occupational and Organizational Psychology, 86, 264282.
Harrison D. A., & Klein, K. J. (2007). What’s the difference? Diversity constructs as separation,
variety, or disparity in organizations. Academy of Management Review, 32, 1199-1228.

104

Hebl, M., & Avery, D. R. (2012). Diversity in organizations. In Weiner, I.B., Schmitt, N.W. &
Highhouse, S., Handbook of Psychology, Vol 12: Industrial and Organizational
Psychology, (pp. 677-697). Hoboken, NJ: John Wiley & Sons.
Hewstone, M., Martin, R., Hammer-Hewstone, C., Crisp, R. J., & Voci, A. (2001). Minoritymajority relations in organizations: challenges and opportunities. In Social Identity
Processes in Organizational Contexts, ed. MA Hogg, DJ Terry, (pp. 67-86). Philadelphia,
PA: Psychological Press.
Herdman, A. O., & McMillan-Capehart, A. (2010). Establishing a diversity program is not
enough: Exploring the determinants of diversity climate. Journal of Business and
Psychology, 25, 39-53.
Hofhuis, J., van der Zee, K. I., & Otten, S. (2012). Social identity patterns in culturally diverse
organizations: The role of diversity climate. Journal of Applied Social Psychology, 42,
964-989.
Hoppe, A., Fujishiro, K., & Heaney, C. (2013). Workplace racial/ethnic similarity, job
satisfaction, and lumbar back health among warehouse workers: Asymmetric reactions
across racial/ethnic groups. Journal of Organizational Behavior, 35, 172-193.
Hoobler, J. M., Lemmon, G., & Wayne, S. J. (2014). Women’s Managerial Aspirations An
Organizational Development Perspective. Journal of Management, 40, 703-730.
Huffman, M. L. (2013). Organizations, managers, and wage inequality. Sex Roles, 68, 216-222.
Huffman, M. L., Cohen, P. N. & Pearlman, J. (2010). Engendering change: organizational
dynamics and workplace gender desegregation, 1975-2005. Administrative Science
Quarterly 55, 255-277.
Ibarra, H. (1992). Homophily and differential returns: Sex differences in network structure and
access in an advertising firm. Administrative Science Quarterly, 37, 422-447.
Jackson, E. S., Brett, J. F., Sessa, V. I., Cooper, D. M., Julin, J. A., & Peyronnin, K. (1991).
Some differences make a difference: individual dissimilarity and group heterogeneity as
correlates of recruitment, promotions, and turnover. Journal of Applied Psychology, 76,
675-689.
Jeanquart-Barone, S. (1996). Implications of racial diversity in the supervisor-subordinate
relationship. Journal of Applied Social Psychology, 26, 935-944.
Joshi A., Liao H., Jackson, S. E. (2006). Cross-level effects of workplace diversity on sales
performance and pay. Academy of Management Journal, 49, 459-481.
Joshi, A., Liao, H., & Roh, H. (2011). Bridging domains in workplace demography research:
A review and reconceptualization. Journal of Management, 37, 521-552.
105

Kalinoski, Z. T., Steele‐Johnson, D., Peyton, E. J., Leas, K. A., Steinke, J., & Bowling, N. A.
(2013). A meta‐analytic evaluation of diversity training outcomes. Journal of
Organizational Behavior, 34, 1076-1104.
Kaplan, D. M., Wiley, J. W., & Maertz, C. P. (2011). The role of calculative attachment in the
relationship between diversity climate and retention. Human Resource Management, 50,
271-287.
Kanter, R. M. (1977a). Men and Women of the Corporation. New York: Basic Books.
Kanter, R. M. (1977b). Some effects of proportions on group life: Skewed sex ratios and
responses to token women. American Journal of Sociology, 82, 965-990.
King, E. B., Hebl, M. R., George, J. M., & Matusik, S. F. (2010). Understanding tokenism:
Antecedents and consequences of psychological climate for gender inequity. Journal of
Management, 36, 537-554.
Konrad, A., & Gutek, B. A. (1987). Theory and research on group composition: Applications to
the status of women and minorities. In S.Oskamps and S. Spacapan (Eds.), Interpersonal
processes, (pp. 85-113).
Konrad, A. M., Winter, S., & Gutek, B. A. (1992). Diversity in work group sex composition:
Implications for majority and minority members. In Tolbert, P. & Bachrach, S. B. (Eds),
Research in the Sociology of Organizations, 10. Greenwich, CT: JAI Press, 115-140.
Kossek, E. E., & Zonia, S. C. (1993). Assessing diversity climate: A field study of reactions to
employer efforts to promote diversity. Journal of Organizational Behavior, 14, 61-61.
Kraiger, K., & Ford, J. K. (1985). A meta analysis of ratee race effects in performance
evaluations. Journal of Applied Psychology, 70, 56-65.
Lau, D. C., & Murnighan, J. K. (1998). Demographic diversity and faultlines: 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.
Liao, H., Joshi, A., & Chuang, A. (2004). Sticking out like a sore thumb: Employee dissimilarity
and deviance at work. Personnel Psychology, 57, 969-1000.
Luo, W. & Azen, R. (2012). Determining predictor importance in hierarchical linear models
using dominance analysis. Journal of Educational and Behavioral Statistics, 38, 3-31.
Madera, J. M., Dawson, M., & Neal, J. A. (2013). Hotel managers’ perceived diversity climate
and job satisfaction: The mediating effects of role ambiguity and conflict. International
Journal of Hospitality Management, 35, 28-34.
106

Magee, J. C., & Galinsky, A. D. (2008). Social hierarchy: The self- reinforcing nature of power
and status. Academy of Management Annals, 2, 351-398.
McGinn, K. L., & Milkman, K. L. (2013). Looking up and looking out. Career mobility effects
of demographic similarity among professionals. Organizational Science, 24, 1041-1060.
McKay, P. F., Avery, D. R., Liao, H., & Morris, M. A. (2011). Does diversity climate lead to
customer satisfaction? It depends on the service climate and business unit demography.
Organization Science, 22, 788-804.
McKay, P. F., Avery, D. R., & Morris, M. A. (2009). A tale of two climates: Diversity climate
from subordinates’ and managers’ perspectives and their role in store unit sales
performance. Personnel Psychology, 62, 767-791.
McKay, P. F., Avery, D. R., & Morris, M. A. (2008). Mean racial-ethnic differences in employee
sales performance: the moderating role of diversity climate. Personnel Psychology, 61,
349-374.
McKay, P. F., Avery, D. R., Tonidandel, S., Morris, M. A., Hernandez, M., & Hebl, M. R.
(2007). Racial differences in employee retention: Are diversity climate perceptions the
key? Personnel Psychology, 60, 35-62.
McPherson, J. M., Smith-Lovin L, Cook, J. M. (2001). Birds of a feather: homophily in social
networks. Annual Review of Sociology, 27, 415-44.
Mehra, A., Kilduff, M., & Brass, D. J. (1998). At the margins: A distinctiveness approach to the
social identity and social networks of underrepresented groups. Academy Management
Journal, 41, 441-452.
Moch, M. K. (1980). Racial differences in job-satisfaction: Testing 4 common explanations,”
Journal of Applied Psychology, 65, 1980, 299-306.
Mor Barak, M. E., Cherin, D. A., & Berkman, S. (1998). Organizational and personal
dimensions in diversity climate: Ethnic and gender differences in employee diversity
perceptions. Journal of Applied Behavioral Sciences, 34, 82-104.
Muthén, L. K., & Muthén, B. O. (1998–2010). Mplus User’s Guide. Sixth Edition. Los
Angeles, CA: Muthén & Muthén.
Nishii, L. (2013). The benefits of climate for inclusion for gender diverse groups. Academy of
Management Journal, 56, 1754-1774.
Oldmeadow, J. A., Platow, M. J., Foddy, M., & Anderson, D. (2003). Self-categorization, status
and social influence. Social Psychology Quarterly, 66, 138-152.

107

Ostroff, C., Kinicki, A.J. & Muhammad, R.S. (2012) Organizational culture and climate. In
Weiner, I.B., Schmitt, N.W. & Highhouse, S., Handbook of Psychology, Vol 12:
Industrial and Organizational Psychology, 643-676. Hoboken, NJ: John Wiley & Sons.
Pfeffer J. (1991). Organization theory and structural perspectives on management. Journal of
Management, 17, 789-803.
Pfeffer, J. (1983). Organizational demography. In L. L. Cummings & B. M. Staw (Eds.),
Research in organizational behavior: 299-357, Greenwich, CT: JAI.
Podsakoff, P. M., MacKenzie, S. B., Lee, J., & Podsakoff, N. P. (2003). Common method biases
in behavioral research: A critical review of the literature and recommended remedies.
Journal of Applied Psychology, 88, 879-903.
Pugh, S. D., Dietz, J., Brief, A. P., & Wiley, J. W. (2008). Looking inside and out: The impact of
employee and community demographic composition on organizational diversity climate.
Journal of Applied Psychology, 93, 1422-14428.
Richard, O. C. (2000). Racial diversity, business strategy, and firm performance: A resourcebased view. Academy of Management Journal, 43, 164-177.
Richard, O. C., McMillan, A., Chadwick, K., & Dwyer, S. (2003). Employing an innovation
strategy in racially diverse workforces: Effects on firm performance. Group &
Organization Management, 28, 107-126.
Ragins, B. R. (1997). Diversified mentoring relationships in organizations: A power perspective.
Academy of Management Review, 22, 482-521.
Reskin, B. F. & McBrier, D. B., (2000). Why not ascription? Organizations, employment of male
and female managers. American Sociological Review 65, 210-33.
Reskin, B. F., McBrier, D. B., & Kmec, J. (1999). The determinants and consequences of
workplace sex and race composition. Annual Review of Sociology, 25, 355-361.
Ridgeway, C. L. (2014). Why status matters for inequality. American Sociological Review, 79, 116.
Ridgeway, C. L. (2001a). Social Status and Group Structure. In Blackwell Handbook of Social
Psychology: Group Processes, edited by Michael A. Hogg and Scott Tindale, 352–75.
Malden, Mass.: Blackwell.
Ridgeway, C. L. (2001b). Gender, status, and leadership. Journal of Social Issues, 57, 637-656.
Riordan, C. M., & Shore, L. M. (1997) Demographic diversity and employee attitudes: An
empirical examination of relational demography within work units. Journal of Applied
Psychology, 82, 342-358.
108

Sacco, J. M., Scheu, C. R., Ryan, A. M., Schmitt, N. (2003). An investigation of race and sex
similarity effects in interviews: A multilevel approach to relational demography. Journal
of Applied Psychology, 88, 852-865.
Sackett, P. R., & Larson, J. R., Jr. (1990). Research strategies and tactics in industrial and
organizational psychology. In M. D. Dunnette & L. M. Hough (Eds.), Handbook of
industrial and organizational psychology (2nd ed., Vol. 1, pp. 419- 489). Palo Alto, CA:
Consulting Psychologists Press.
Schneider, B., Ehrhart, M. G., & Macey, W. H. (2013). Organizational climate and culture.
Annual Review of Psychology, 64, 361-388.
Schneider, B., Macey, W. H., Lee, W., & Young, S. A. (2009). Organizational service climate
drivers of the American Customer Satisfaction Index (ACSI) and financial and market
performance. Journal of Service Research, 12, 3-14.
Sellers, R. M., Smith, M. A., Shelton, J. N., Rowley, S. A. J., & Chavous, T. M. (1998).
Multidimensional model of racial identity: A reconceptualization of African American
racial identity. Personality and Social Psychology Review, 2, 18-39.
Skaggs, S., & DiTomaso, N. (2004). Understanding the effects of workforce diversity on
employment outcomes: A multidisciplinary and comprehensive framework. Research in
the Sociology of Work, 14, 279-306.
Shore, L. M., Randel, A. E., Chung, B. G., Dean, M. A., Ehrhart, K. H. & Singh, G. (2011).
Inclusion and diversity in work groups: A review and model for future research. Journal
of Management, 37, 1262-1289.
Siemsen, E., Roth, A. & Oliveira, P. (2010). Common method bias in regression models with
linear, quadratic, and interaction effects, Organizational Research Methods, 13, 456-476.
Smith, P. L., Smits, S. J., & Hoy, F. (1998). Employee work attitudes: The subtle influence of
gender. Human Relations, 51, 649-666.
Snijders, T., & Bosker, R. J. (1994). Modeled variance in two-level models. Sociological
Methods & Research, 22, 342-363.
Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation
models. Sociological Methodology, 13, 290-312.
Tajfel, H. (1982). Social psychology of intergroup relations. Annual Review of Psychology, 33,
1-39.
Thye, S. R. (2000). A status value theory of power in exchange relations. American Sociological
Review, 65, 407-432.
Tsui, A. S., Egan, T. D., & O’Reilly, C. A. (1992). Being different: Relational demography and
109

organizational attachment. Administrative Science Quarterly, 37, 549-579.
Tsui, A. S., & Gutek, B. A. (1999). Demographic Differences in Organizations. Oxford:
Lexington Books.
Tsui, A., & O’Reilly, C. (1989). Beyond simple demographic effects: The importance of
relational demography in superior-subordinate dyads. Academy of Management Journal,
32, 402-423.
Tajfel, H., & Turner, J. (1986). The social identity theory of intergroup behavior. In S. Worchel
& W. Austin (Eds.), Psychology of intergroup relation, 7-24. Chicago: Nelson-Hall.
Tonidandel, S., Avery, D., Bucholtz, B., & McKay, P. (2008). An alternative explanation for the
asymmetrical effects in relational demography research. Personnel Psychology, 61, 617633.
Van Dijk, H., van Engen, M. L., & van Knippenberg, D. (2012). Defying conventional wisdom:
A meta-analytical examination of the differences between demographic and job-related
diversity relationships with performance. Organizational Behavior and Human Decision
Processes, 119, 38-53.
Vecchio, R. P., & Bullis, R. C. (2001). Moderators of the influence of supervisor–subordinate
similarity on subordinate outcomes. Journal of Applied Psychology, 86, 884-896.
Williams, K. Y. & O’Reilly, C. A., III. (1998). Demography and diversity in organizations: A
review of 40 years of research. Research in Organizational Behavior, 20. Greenwich,
CT: JAI Press, 77–140.
Winfield. I., & Rushing, B. (2005). Bridging the border between work and family: The effects of
supervisor-employee similarity. Sociological Inquiry, 75, 55-80.
Wharton, A. S., & Baron, J. N. (1991). Satisfaction? The psychological impact of gender
segregation on women at work. The Sociological Quarterly, 32, 365-387.
Wolfson, N., Kraiger, K., Finkelstein, L. (2011). The relationship between diversity climate
perceptions and workplace attitudes. The Psychologist-Manager Journal, 14, 161-176.
Zhang, Z., Zyphur, M. J., & Preacher, K. J. (2009). Testing multilevel mediation using
hierarchical linear models: Problems and solutions. Organizational Research Methods
12, 695-719.
Zatzick, C. D., Elvira, M. M., & Cohen, L. E. (2003) When is more better? The effects of racial
composition on voluntary turnover. Organizational Science, 14, 483-496.

110