Journal of Occupational Health Psychology © 2010 American Psychological Association

2010, Vol. 15, No. 2, 191–207 1076-8998/10/$12.00 DOI: 10.1037/a0018756

A Multilevel Model of the Effects of Equal Opportunity Climate

on Job Satisfaction in the Military
Benjamin M. Walsh Russell A. Matthews
University of Connecticut Louisiana State University

Michael D. Tuller Kizzy M. Parks

University of Connecticut K. Parks Consulting, Inc., Melbourne, FL
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Daniel P. McDonald
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Defense Equal Opportunity Management Institute, Patrick Air Force Base, FL

To date, minimal work has explored associations between equal opportunity (EO) climate and
employee work attitudes, and no known research has investigated the effects of EO climate
beyond the individual level. We address these gaps in the literature by testing a multilevel
structural equation model in which effects of EO climate are considered at both the individual and
unit levels. At the individual level, we predicted that psychological EO climate would be directly
associated with job stress and job satisfaction, as well as indirectly related to job satisfaction via
stress. In addition, cross-level associations between unit EO climate and job stress and job
satisfaction were hypothesized to be mediated by cohesion. Findings supported the proposed
model; hypothesized relations were supported at both levels of analysis. We conclude with a
discussion of the findings, study limitations, and directions for future EO climate research.

Keywords: equal opportunity climate, cohesion, job satisfaction, job stress, multilevel structural
equation modeling

With women and racial/ethnic minorities accounting
for a large portion of the United States military popu-
lation, the military has committed itself to continuing to
provide equal opportunity (EO) for all personnel to
create a force that is representative of the people it
serves (Edwards, 2001). Over the past two decades, the
assessment of perceptions of discriminatory employ-
ment practices in the military has become a frequent
Benjamin M. Walsh and Michael D. Tuller, Department
of Psychology, University of Connecticut; Russell A. Mat-
thews, Department of Psychology, Louisiana State Univer-
sity; Kizzy M. Parks, K. Parks Consulting, Inc., Melbourne,
FL; and Daniel P. McDonald, Defense Equal Opportunity
Management Institute, Patrick Air Force Base, FL.
The work of Benjamin M. Walsh and Michael D. Tuller
was supported in part by Grant No. 5 T01 OH008610-04
from CDC-NIOSH. The views expressed in this paper do
not necessarily reflect those of NIOSH, the United States
Department of Defense, the Defense Equal Opportunity
Management Institute, or any other federal agencies. An
earlier version of this research was presented at the 24th
annual meeting of the Society for Industrial and Organiza-
tional Psychology in New Orleans, LA.
Correspondence concerning this article should be ad-
dressed to Benjamin M. Walsh, Department of Psychology,
University of Connecticut, 406 Babbidge Road, Unit 1020,
Storrs, CT 06269. E-mail: benjamin.walsh@uconn.edu
practice (e.g., Dansby & Landis, 1991; Newell, Rosen-
feld, & Culbertson, 1995; McIntyre, Bartle, Landis, &
Dansby, 2002; Rosenfeld, Newell, & Le, 1998) with
researchers focusing primarily on perceptions of equal
opportunity climate (EO climate; Dansby & Landis,
1991). However, little empirical research has investi-
gated the process through which EO climate affects
employees’ work attitudes, and to our knowledge no
research has studied EO climate as a unit level construct
reflecting the shared climate perceptions of unit mem-
bers.
In the present study, we address these gaps in the
literature by examining EO climate at both the individ-
ual and unit levels of analysis. At the individual level,
we study EO climate as a predictor of job stress and job
satisfaction. At the unit level, EO climate is modeled as
a direct predictor of unit cohesion, a variable which has
beneficial effects on individual satisfaction, perfor-
mance, and mission effectiveness (Oliver, Harman,
Hoover, Hayes, & Pandhi, 1999). Unit EO climate is
also hypothesized to have significant cross-level indi-
rect effects on individuals’ job stress and job satisfac-
tion. In the following sections, we describe in greater
detail the construct of EO climate, ground EO climate
within the general climate literature, and briefly sum-

191
 192 WALSH, MATTHEWS, TULLER, PARKS, AND MCDONALD
marize previous EO climate research. We then outline
our conceptual model, drawing on fairness theories and
models of job stress as rationale for our hypotheses.
Equal Opportunity Climate
We define EO climate as employees’ perceptions
of the degree to which discrimination and harassment
are likely to occur within their work unit (Dansby &
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
Landis, 1991). EO climate pertains specifically to
This document is copyrighted by the American Psychological Association or one of its allied publishers.
discrimination and harassment surrounding one or
more legally protected categories (e.g., race, reli-
gion). Perceptions of a positive EO climate indicate
that these forms of mistreatment are unlikely to oc-
cur, whereas perceptions of a negative EO climate
suggest that discrimination and harassment are likely,
thereby reflecting a work environment characterized
by unfair and disrespectful behavior.
Researchers have operationalized EO climate in
terms of several forms of harassment and discrimi-
nation including sexual and racial harassment, and
discrimination on the basis of disability, age, and
religion (Dansby & Landis, 1991; Landis, Dansby, &
Faley, 1993; Truhon & Parks, 2007). This operation-
alization is consistent with armed services law in-
cluding Executive Order No. 9981 (1948) which
mandates equal treatment of all persons in the mili-
tary, regardless of race, color, religion, or national
origin. In the current study, we focus our assessment
on one form of harassing behavior, racist behavior,
and four forms of discrimination, including race, age,
religious, and disability discrimination (Dansby &
Landis, 1991).1 This focus is congruent with previous
research and also aligns with the research goal of the
Defense Equal Opportunity Management Institute
(DEOMI) which conducts research on the degree to
which EO climate contributes to organizational ef-
fectiveness (DEOMI, 2007).
EO climate is traditionally conceptualized as a
psychological climate as opposed to a unit or orga-
nizational climate (James & Jones, 1974). Schneider
and Reichers (1983) distinguished between psycho-
logical and organizational climate in terms of the unit
of analysis. Whereas psychological climate refers to
individual perceptions of the work context, organiza-
tional climate reflects shared climate perceptions ag-
gregated to the unit or organization level. Further-
more, climate researchers (e.g., Rousseau, 1988;
Schneider & Reichers, 1983) argue for the assess-
ment of specific climates pertaining to a particular
domain (e.g., climate for safety, Zohar, 1980; climate
for service, Schneider, White, & Paul, 1998), thereby
asserting that multiple climates exist within any par-
ticular work environment.
As mentioned above, to our knowledge, EO cli-
mate has been studied only at the individual level as
psychological EO climate, without considering the
possibility that employees working in the same con-
text might have shared perceptions of the EO climate
within their work environment. Although EO climate
perceptions are formed at the individual level and it is
important to consider effects of psychological EO
climate, we suggest that EO climate is also a mean-
ingful unit-level construct. Because individuals
working in the same location are exposed to similar
policies, practices, procedures, and behavioral norms
with respect to EO, we believe that agreement in EO
climate perceptions will be observed among unit
members (Schneider & Reichers, 1983). More spe-
cifically, psychological EO climate perceptions are
expected to become shared EO climate perceptions
over time in a bottom-up process of emergence (Koz-
lowski & Klein, 2000). Psychological EO climate
perceptions are theorized to manifest as unit EO
climate through a direct consensus composition
model such that the meaning of unit EO climate is in
the consensus among unit members (Chan, 1998).
Similar constructs pertaining to organizational fair-
ness perceptions have also been conceptualized as
emerging at higher levels beyond the individual em-
ployee (e.g., justice climate; Colquitt, Noe, & Jack-
son, 2002; Liao & Rupp, 2005; Naumann & Bennett,
2000). Accordingly, in the current study we consider
the influence of EO climate simultaneously at both
the individual and unit levels.
EO climate is often seen as interchangeable with
diversity climate despite the fact that they have fun-
damental differences (Marquis, Lim, Scott, Harrell,
& Kavanagh, 2008). Consistent with the definition
outlined above, diversity researchers commonly de-
fine EO climate in terms of demographic diversity,
policy and/or fairness (Hicks-Clarke & Iles, 2000;
Kossek & Zonia, 1993). EO is legally mandated and
requires compliance to ensure that organizations do
not engage in discriminatory employment practices
or allow a hostile work environment to foster. In the
1
As noted, previous EO climate research has investigated
additional forms of harassing behaviors, such as sexual
harassment and sex-related discrimination. The addition of
harassment and discrimination of a sexual nature in the
current study would have helped to create a more robust and
well-rounded perspective of EO climate. However, we did
not have access to information with respect to sexual ha-
rassment or sex discrimination for use in the current study.
 EQUAL OPPORTUNITY CLIMATE 193

military, each service has an EO office to ensure that
personnel are treated with dignity and respect. These
offices are supported by a multimillion dollar budget
which funds proactive efforts including preventive ra-
cial harassment training and complaint investigations
(Edwards, 2001). Conversely, diversity climate is de-
fined as individual, unit, or organizational perceptions
of how employees feel their organization values diver-
sity and maximizes the potential advantages of the
workforce (Cox, 1993; Mor Barak, Cherin, & Beck-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
cently some researchers have explored relations be-
tween EO climate and work-related attitudes. For
example, McIntyre et al. (2002) found support for a
model linking EO climate to perceptions of work
group efficacy, job satisfaction, and organizational
commitment. It is interesting that McIntyre et al.
(2002) distinguished between EO climate perceptions
directed at the broader organization and more specific
work group EO climate perceptions. However, their
conceptualization of EO climate was limited in scope

man, 1998). Thus, diversity climate is strategic in focus because it focused almost exclusively on race dis-
This document is copyrighted by the American Psychological Association or one of its allied publishers.

and reflects perceptions of initiatives aimed at capital-
izing on the strengths and minimizing the potential
weaknesses of employing a demographically diverse
workforce (Cox, 1993; Mor Barak et al., 1998).
The focus of a great deal of EO climate research is
on the psychometric properties of EO climate mea-
sures (Dansby & Landis, 1991; Estrada, Stetz, &
Harbke, 2007; Landis et al., 1993; Truhon, 2008) and
group differences in EO climate perceptions and EO
climate– outcome relationships (Dansby & Landis,
1998; Estrada & Harbke, 2008; Truhon, 2008). Re-
crimination. In addition, all constructs were studied
at the individual level, despite incorporating unit-
level constructs into the study. Finally, McIntyre et
al. (2002) did not investigate explanatory variables
linking EO climate to work attitudes. Although sup-
port was found for the direct effect of EO climate, it
is plausible that additional process (i.e., mediating)
variables might better explain how EO climate af-
fects attitudinal outcomes.
To that end, we turn to a brief discussion of our
conceptual multilevel model presented in Figure 1.

Figure 1. Conceptual multilevel model of outcomes associated with EO climate. RB ⫽

racist behavior; RD ⫽ race discrimination; AD ⫽ age discrimination; RLD ⫽ religious
discrimination; and DD ⫽ disability discrimination. In the within-unit model, the filled circles
indicate that these variables have random intercepts. In the between-unit model, these random
intercepts are latent variables which vary across units (Muthén & Asparouhov, 2009).
 194 WALSH, MATTHEWS, TULLER, PARKS, AND MCDONALD
The model specifies effects of EO climate at the
individual (i.e., the within-unit model, below the dot-
ted line in Figure 1) and unit (i.e., the between-unit
model, above the dotted line in Figure 1) levels. At
the individual level, psychological EO climate is
posited to have direct effects on job stress and job
satisfaction, as well as an indirect effect on job sat-
isfaction via job stress. Further, cohesion is modeled
as an explanatory variable at the unit level linking
unit EO climate to job stress and job satisfaction. The
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
indirect effects of unit EO climate on job stress and
This document is copyrighted by the American Psychological Association or one of its allied publishers.
job satisfaction via cohesion represent forms of
cross-level mediation (Mathieu & Taylor, 2007).
These cross-level relationships are presented in Fig-
ure 1 at the unit level because, from an analytical
perspective, it is the unit-level variance in job stress
and job satisfaction that is affected by unit EO cli-
mate and cohesion.2 In the sections that follow, we
provide the theoretical and empirical rationale for our
hypotheses.
Underlying Theoretical Rationale and
Study Hypotheses for Effects of
Psychological EO Climate
Research on fairness in organizations offers a
guiding framework for conceptualizing effects of EO
climate at multiple levels of analysis. This is because
EO climate reflects broad EO-related fairness percep-
tions including aspects of distributive (i.e., fairness in
the distribution of assignments, promotions, and
other work opportunities; Adams, 1965), procedural
(i.e., fairness in the processes used to determine out-
comes; Lind & Tyler, 1988), and interactional justice
(i.e., fair and respectful treatment by supervisors or
other authority figures; Bies & Moag, 1986), as well
as more general fair and respectful treatment among
employees (see Table 1 for EO climate items by
factor). The evidence offered by Colquitt, Conlon,
Wesson, Porter, and Ng’s (2001) meta-analysis pro-
vides clear support for the positive relationship be-
tween work-related fairness perceptions and work
attitudes including job satisfaction. Given this meta-
analytic evidence, it is not surprising that studies
consistently indicate direct effects between psycho-
logical EO climate and job satisfaction, such that
individuals holding more positive EO climate percep-
tions also report greater satisfaction with their jobs
(e.g., Estrada & Harbke, 2008; McIntyre et al., 2002).
Based on this research, we expect that psychological
EO climate is positively related to job satisfaction.
Hypothesis 1: Psychological EO climate is pos-
itively related to job satisfaction.
We also propose that psychological EO climate
has direct effects on job stress, which is defined in the
present study as job-related anxiety and tension
(Parker & DeCotiis, 1983). To our knowledge, the
direct relationship between psychological EO climate
and stress has not been examined. However, it seems
intuitive that employees who perceive a fair work
environment, one in which harassing and discrimina-
tory behavior is unlikely to occur, will also report
lower levels of job-related stress. Indeed, theory and
research support such a link. For example, Judge and
Colquitt (2004) developed a strong argument for
what they termed the “injustice as stressor perspec-
tive” (p. 395) which suggests that perceived unfair-
ness may act as a kind of job-related stressor. They
based their argument on the conceptual overlap
among justice constructs and common work stressors
(e.g., interpersonal conflict at work; Spector & Jex,
1998), and relevant theories of justice and fairness. In
particular, Judge and Colquitt (2004) studied the ef-
fects of justice perceptions on stress, as well as the
extent to which work-family conflict mediates the
justice—stress relation. Judge and Colquitt (2004)
observed significant negative correlations between
several justice dimensions and stress, and found sig-
nificant indirect effects of procedural and interper-
sonal justice on stress via work-family conflict, thus
2
Because of the two-level nature of the data, the variance
in the variables modeled at both levels (i.e., EO climate, job
stress, job satisfaction) is decomposed into two latent com-
ponents, a between-unit component (i.e., the variance be-
tween-units) and a within-unit component (i.e., the variance
within-units; Asparouhov & Muthén, 2006). We hypothe-
size that unit EO climate affects individual’s job satisfaction
and job stress via cohesion (i.e., a cross-level indirect ef-
fect), but it is important to note that it is the between-unit
variance in job satisfaction and job stress that is explainable
by unit EO climate and cohesion; the individual-level (with-
in-unit) variance in job stress and satisfaction are assumed
to be unaffected by unit-level variables (Preacher, Zyphur,
& Zhang, in press). This is why job stress and job satisfac-
tion are represented in Figure 1 at both the unit and indi-
vidual levels. Conceptually, we are interested in effects of
unit EO climate and cohesion on individuals’ job stress and
satisfaction, not units’ job stress and satisfaction. To appro-
priately study unit job stress or job satisfaction, a unit
referent would be needed for all items rather than an indi-
vidual referent such as that used in the current study (see
Chan, 1998). Thus, we refer to job stress and job satisfaction
when discussing these constructs at both levels of analysis
rather than referring to unit versus individual job satisfaction.
 EQUAL OPPORTUNITY CLIMATE 195

Table 1
EO Climate Items Organized by Factor
Item Factor
A person told several jokes about a particular race/ethnicity. Racist behavior
Offensive racial/ethnic names were frequently heard. Racist behavior
Racial/ethnic jokes were frequently heard. Racist behavior
A supervisor did not select for promotion a qualified subordinate of a different race/
ethnicity. Race discrimination
Members of a particular race/ethnicity were assigned less desirable office space than
members of a different race/ethnicity. Race discrimination
The person in charge of the organization changed the duty assignments when it was
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

discovered that two people of the same race/ethnicity were assigned to the same
This document is copyrighted by the American Psychological Association or one of its allied publishers.

sensitive area on the same shift. Race discrimination

While speaking to a group, the person in charge of the organization took more time
to answer questions from one race/ethnic group than from another group. Race discrimination
A younger person was selected for a prestigious assignment over an older person
who was equally, if not slightly better qualified. Age discrimination
An older individual did not get the same career opportunities as did a younger
individual. Age discrimination
A young supervisor did not recommend promotion for a qualified older worker. Age discrimination
A well-qualified person was denied a job because the supervisor did not like the
religious beliefs of the person. Religious discrimination
A demeaning comment was made about a certain religious unit. Religious discrimination
A supervisor favored a worker who had the same religious beliefs as the supervisor. Religious discrimination
A worker with a disability was not given the same opportunities as other workers. Disability discrimination
A career opportunity speech to a worker with a disability focused on the lack of
opportunity elsewhere; to others, it emphasized promotion. Disability discrimination
A supervisor did not appoint a qualified worker with a disability to a new position,
but instead appointed another, less qualified worker. Disability discrimination

providing evidence supporting the link between psy-
chological EO climate and job stress.
Research on a conceptually similar climate con-
struct, violence climate (Kessler, Spector, Chang, &
Parr, 2008; Spector, Coulter, Stockwell, & Matx,
2007), further supports the proposed link between
psychological EO climate and job stress. Violence
climate is defined as “employees’ perceptions of or-
ganizational policies, practices, and procedures re-
garding the control and elimination of workplace
violence and verbal aggression” (Kessler et al., 2008,
p. 110). Thus, violence climate and EO climate have
a similar focus on fair and respectful treatment in the
workplace. Of direct relevance to the present study,
Kessler et al. (2008) found that violence climate
accounted for incremental variance in self-reported
anxiety beyond the effects of personal experiences of
workplace aggression. Consequently, we feel there is
sufficient justification for predicting a direct negative
relationship between psychological EO climate and
job stress.
Hypothesis 2: Psychological EO climate is neg-
atively related to job stress.
Although job stress is a consequence of exposure to
stressors at work, it is also a potential precursor to
additional negative outcomes. Drawing from theory and
empirical findings suggesting that stress resulting from
regular exposure to stressors can ultimately lead to more
distal outcomes (e.g., Lazarus, 1981; Selye, 1976),
Parker and DeCotiis (1983) hypothesized that outcomes
of job stress include decreased organizational commit-
ment and job satisfaction. Others have developed and
tested similar models of antecedents and outcomes of
job stress among workers in various professions (e.g.,
Parasuraman & Alutto, 1984; Summers, DeCotiis, &
DeNisi, 1995) with findings supporting the link be-
tween job stress and work-related attitudes across con-
texts. For instance, Parasuraman and Alutto (1984)
identified negative associations among job stress, and
organizational commitment and job satisfaction for
employees of a food processing plant, and job stress
was a significant predictor of job satisfaction in re-
gression analyses. Similar relationships were re-
ported by Summers et al. (1995) in a sample of
managers of a restaurant chain. Based on these con-
sistent findings with respect to the job stress—job
 196 WALSH, MATTHEWS, TULLER, PARKS, AND MCDONALD
satisfaction relationship, we predict a negative asso-
ciation between job stress and job satisfaction.
Hypothesis 3: Job stress is negatively related to
job satisfaction.
Hypothesis 1 predicts a direct relationship between
psychological EO climate and job satisfaction, Hy-
pothesis 2 predicts a direct relationship between psy-
chological EO climate and job stress, and Hypothesis
3 predicts a direct relationship between job stress and
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
job satisfaction. Together, this collection of hypoth-
eses specifies a partially mediated model between
psychological EO climate and job satisfaction
(Mathieu & Taylor, 2006). Thus, we expect job stress
to partially mediate the relationship between psycho-
logical EO climate and job satisfaction.
Hypothesis 4: The relationship between psycho-
logical EO climate and job satisfaction is par-
tially mediated by job stress.
Underlying Theoretical Rationale and
Study Hypotheses for Effects of Unit
EO Climate
Fairness theory and research is also useful for
hypothesizing effects of unit EO climate. Referent
cognitions theory (Folger, 1987, 1993), and more
recently, fairness theory (Folger & Cropanzano,
2001), posit that individuals evaluate their treatment
in comparison to some referent other, and it is the
comparative evaluations that drive perceptions of fair
or unfair treatment. Fairness theory makes percep-
tions of accountability more salient than referent cog-
nitions theory, wherein individuals hold others ac-
countable for fair (by ascribing credit) or unfair
treatment (by ascribing blame) when others’ behavior
is deemed discretionary. As noted in these theories,
others are often the cause of evaluations of (un)just
treatment, as well as the focus of those evaluations.
Whereas fairness theory pertains to specific events in
the workplace (Folger & Cropanzano, 2001), fairness
heuristic theory (Lind, 2001) focuses on overall fair-
ness perceptions and is thus more consistent with the
nature of EO climate.
Fairness heuristic theory suggests that people use a
fairness heuristic, comprised of general perceptions
of fairness in the workplace, to guide whether efforts
are focused on meeting self interests or others’ inter-
ests (e.g., the interests of one’s unit or organization;
Lind, 2001). When individuals hold positive fairness
perceptions, theory suggests such individuals become
more other-focused and willing to cooperate with and
help others. On the other hand, individuals with neg-
ative fairness perceptions become more self-focused
and concerned with fulfilling personal work needs.
Generalizing to the unit level, it follows that when
unit members have shared positive fairness perceptions,
the unit should be more concerned with ensuring that
unit needs, rather than individual needs, are met. Ulti-
mately, fairness heuristic theory suggests that social
constructs are strongly associated with fairness percep-
tions. As Lind (2001) argues, “the strongest effects of
justice judgments are relational, in that they are conse-
quences that are linked to what early social psycholo-
gists termed unit cohesion and promotive social pro-
cess” (p. 58). Because EO climate reflects overall EO
fairness perceptions, we argue that the most salient
outcomes of unit EO climate are those of an interper-
sonal nature including unit cohesion.
We propose that unit EO climate is an important
driver of cohesion. In the current study cohesion
refers to the social bond among members and the
extent to which members work well with one another.
As fairness heuristic theory suggests, units holding
positive EO climate perceptions should be more will-
ing to cooperate with one another as positive fairness
perceptions are thought to shift members’ focus from
serving self interests to the interests of the unit (Lind,
2001). This shift occurs because positive fairness
perceptions encourage the development of a stronger
social identity with the unit, and greater feelings that
contributions to the unit will be compensated with
fair rewards (Lind, 2001). Thus, we expect a positive
relationship between unit EO climate and cohesion.
Hypothesis 5: Unit EO climate is positively re-
lated to cohesion.
Previous meta-analytic work with military personnel
supports the positive associations among cohesion and
performance, job and military satisfaction, well-being,
retention, and readiness (Oliver et al., 1999). The
positive findings appear to generalize across levels of
analysis. For example, Gully, Devine, and Whitney
(1995) conducted a meta-analysis and found signifi-
cant positive associations between cohesion and per-
formance at both the group and individual levels of
analysis. The confidence intervals for the cohesion–
performance relationships overlapped across levels,
and the strongest relationships existed for group-level
conceptualizations. In a sample of Air Force military
and civilian personnel, Wech, Mossholder, Steel, and
Bennett (1998) investigated the cross-level impact of
cohesion on individual performance and organiza-
 EQUAL OPPORTUNITY CLIMATE
tional commitment, finding significant positive asso-
ciations between cohesion and each outcome. The
positive findings are not surprising considering that
cohesion entails high levels of attraction among
members and a commitment to fulfilling unit objec-
tives. Based on these findings, we expect a positive
relationship between cohesion and job satisfaction.
Hypothesis 6: Cohesion is positively related to
job satisfaction.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Theory and research point to additional beneficial
outcomes of cohesion including decreased job-
related stress. Drawing on reviews of job-related
stressors at the time (e.g., Cooper & Marshall, 1976),
Parker and DeCotiis (1983) developed a framework
of stressors which incorporates aspects of one’s job,
one’s role, organizational structure, the extent of
one’s developmental opportunities, and the quality of
relationships at work, of which the latter subsumes
cohesion. Immediate outcomes of job stressors in-
clude two forms of job stress: time-related job stress
and job-related anxiety, of which the latter form is the
focus of the present study. Parker and DeCotiis
(1983) tested their model and found partial support
for the link between cohesion and job stress, although
analyses were conducted strictly at the individual
level. Individual perceptions of cohesion were a sig-
nificant, negative predictor of time-related job stress,
but not job-related anxiety, a finding which they
attribute to multicollinearity among the relationship-
oriented predictors. Further, Griffith and Vaitkus
(1999) argue that cohesion serves as a resource sim-
ilar to social support which may directly reduce job-
related strain. Although their discussion centered on
strain, cohesion may be expected to have direct ef-
fects on stress as well given that felt stress is a
precursor to more distal strains (Parker & DeCotiis,
1983). Based on this rationale, we expect cohesion to
be negatively related to job stress.
Hypothesis 7: Cohesion is negatively related to
job stress.
Also, we expect the same negative relationship be-
tween job stress and job satisfaction at the unit level that
is hypothesized at the individual level of analysis, based
on the theory (e.g., Parker & DeCotiis, 1983) and
empirical findings (e.g., Parasuraman & Alutto,
1984; Summers et al., 1995) reviewed previously.
Hypothesis 8: Job stress is negatively related to
job satisfaction.
197
Finally, the relationships specified for unit EO
climate imply fully mediated cross-level associations
between unit EO climate and job stress and unit EO
climate and job satisfaction. Unit EO climate is ex-
pected to indirectly affect job satisfaction via two
pathways: (a) unit EO climate— cohesion—job sat-
isfaction, and (b) unit EO climate— cohesion—job
stress—job satisfaction. We simply note that a fully
mediated relationship between unit EO climate and
job satisfaction is hypothesized. However, cohesion
is the sole mediator of the hypothesized indirect
relationship between unit EO climate and job stress.
In sum, we hypothesize fully mediated relationships
between unit EO climate and job satisfaction and unit
EO climate and job stress.
Hypothesis 9: The relationship between unit EO
climate and job satisfaction is fully mediated by
cohesion and job stress.
Hypothesis 10: The relationship between unit
EO climate and job stress is fully mediated by
cohesion.
Method
Participants
Data used to test the multilevel model were collected
by DEOMI from United States Army personnel who
completed a survey in the spring of 2008. A total of
2,179 respondents completed the survey and we fo-
cused our analysis on Army units who were not de-
ployed outside the continental United States. Personnel
deployed at the time of the survey were excluded be-
cause we anticipated differences between active duty
personnel located in the United States and those de-
ployed overseas due to involvement in wartime opera-
tions such as in Afghanistan and Iraq. For example, the
suicide rate among Army personnel has increased since
the Iraq war, with Army officials pointing to stress from
war operations as a likely cause for this increase (Tyson,
2008). Moreover, Adler, Huffman, Bliese, and Castro
(2005) found a negative association between deploy-
ment length and psychological well-being, as well as
evidence that the effect may be stronger for men than
women. In addition, units with responses from fewer
than three members were excluded from analyses; this
is a common practice in group-level and multilevel
research (e.g., Gilson, Mathieu, Shalley, & Ruddy,
2005; Glomb & Liao, 2003).
Applying the aforementioned criteria resulted in a
final unit-level N ⫽ 66 and an individual-level N ⫽
 198 WALSH, MATTHEWS, TULLER, PARKS, AND MCDONALD

1,394 (Munit size ⫽ 21.12). The majority of the sample

data (N ⫽ 66). Correlations below the diagonal are based on individual-level data and correlations above the diagonal are based on unit-level data. Coefficient alpha reliability
estimates are presented along the diagonal; the estimate on the left is based on individual-level data and the estimate on the right is based on unit-level data (where applicable).
Note. Means and standard deviations without parentheses are based on individual-level data (N ⫽ 1,394) and means and standard deviations in parentheses are based on unit-level
.48ⴱⴱ

.48ⴱⴱ
.38ⴱⴱ
.83ⴱⴱ
⫺.55ⴱⴱ
.30ⴱ

(.85)
.20
8
was White (57.0%) with another 29.4% indicating
they were Black. Approximately 10% of participants
indicated that they were Hispanic. Most respondents

ⴱⴱ

⫺.35ⴱⴱ
⫺.36ⴱⴱ

⫺.45ⴱⴱ
were male (74.4%) and between 20 and 30 years of

⫺.30ⴱ
⫺.28ⴱ

(.87)
⫺.39

⫺.46
7
age (72.4%). Enlisted personnel accounted for a ma-
jority of the sample (74.5%). These sample demo-
graphics are reflective of the composition of active

(.92) (.97)
.38ⴱⴱ

.35ⴱⴱ
.33ⴱⴱ
duty United States Army personnel (Defense Man-

.29ⴱ
.20

⫺.36
.58
6
power Data Center, 2007).
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Procedure

(.87) (.93)
ⴱⴱ

.54ⴱⴱ
.65ⴱⴱ
.58ⴱⴱ
All participants completed the DEOMI Organiza-

.38

.26
⫺.22
.26
5
tional Climate Survey (DEOCS). The DEOCS is
typically deployed annually at the request of a mili-
tary unit commander and is similar to an annual

(.78) (.69)
employee survey. The DEOCS is administered and

ⴱⴱ

.69ⴱⴱ
.54ⴱⴱ
.60

.66
.32
⫺.28
.33
4
received by DEOMI and is available in paper-and-
pencil and web-based versions. All personnel are
provided with either a confidential unique online
code to complete the survey online, or a paper copy

(.90) (.92)
ⴱⴱ

.50ⴱⴱ
of the survey and a response sheet. Participants are .43

.63
.69
.30
⫺.29
.31
3

then able to complete the survey at their own discre-

tion. Survey responses are completely confidential,
and although the DEOCS is deployed at the request
(.86) (.92)

of a military unit commander, the commander does

ⴱⴱ
.37

.58
.61
.59
.35
⫺.25
.33
not receive specific details about individual respon-
2

dents in terms participation or outcomes on the sur-

vey. The DEOCS evolved from the Military Equal
Opportunity Climate Survey (MEOCS; Dansby &
(.86) (.94)

Landis, 1991), and both surveys are suitable for mil-

.49
.44
.55
.46
.37
⫺.31
.31
1

itary and civilian organizations of varying sizes. In

total, the DEOCS contains 66 self-report items which
are traditionally combined into 13 distinct scales, of
which seven address equal employment opportunity
.89 (.45)
.65 (.26)
.80 (.28)
.67 (.21)
.70 (.25)
.96 (.46)
1.16 (.45)
.88 (.38)

All individual-level correlations are significant at p ⬍ .01.

and six address organizational effectiveness factors.

SD

Previous psychometric work conducted on the

DEOCS and the MEOCS provide support for the
Descriptive Statistics for Study Variables

internal consistency and factor structure of the DE-

4.18 (4.18)
4.60 (4.62)
4.49 (4.54)
4.57 (4.61)
4.60 (4.65)
3.98 (3.94)
2.67 (2.67)
3.84 (3.85)

OCS scales (Estrada et al., 2007; Landis, Fisher, &

Mean

Dansby, 1988; Truhon, 2003).

Measures
Disability discrimination
Religious discrimination

Below is a description of the measures used in the

p ⬍ .05. ⴱⴱ p ⬍ .01.
Race discrimination

current study. For all measures, higher scores reflect

Age discrimination

higher levels of each construct (items were recoded

Construct
Racist behavior

Job satisfaction

as necessary for all scales other than EO climate). See

Job stress

Table 2 for internal consistency reliability informa-

Cohesion

tion for all study variables.

Table 2

EO climate. EO climate was assessed as a latent

construct comprised of five categories of discrimina-
1.
2.
3.
4.
5.
6.
7.
8.

ⴱ
 EQUAL OPPORTUNITY CLIMATE
tion and harassment. Specifically, race discrimina-
tion was assessed with four items, while racist be-
havior, age discrimination, religious discrimination,
and disability discrimination were measured with
three items each. Participants evaluated each item in
terms of the likelihood that each behavior could have
occurred at their duty location during the 30 work
days prior to survey administration. All items were
evaluated on a 5-point response scale (1 ⫽ there is a
very high chance that the action occurred, to 5 ⫽
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
there is almost no chance that the action occurred).
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Cohesion. Cohesion was measured with four
items assessing the social bond among members and
how well members come together to complete tasks
(e.g., “My work group works together well as a
team,” “Members of my work group trust each
other”). All items were answered on a 5-point re-
sponse scale (1 ⫽ totally agree with the statement, to
5 ⫽ totally disagree with the statement).
Job stress. Job stress was examined with four of
the five items from Parker and DeCotiis’s (1983)
scale of job stress which was included in the DEOCS
for the current administration. Specifically, items
were included which assessed job-related anxiety
(e.g., “My job gets to me more than it should”).
Although these data were collected by DEOMI and
thus were not collected for use in this particular
study, we suspect that the fifth item in the measure
(i.e., “I feel guilty when I take time off from job”)
was not included because it appears to assess percep-
tions of guilt rather than anxiety. This is substantiated
by the fact that the item had the lowest factor loading
(.51) of the five items reported by Parker and
DeCotiis (loadings for remaining items ranged from
.67–.76). All items were answered on a 5-point re-
sponse scale (1 ⫽ totally agree with the statement, to
5 ⫽ totally disagree with the statement).
Job satisfaction. Job satisfaction was measured
with five items assessing one’s satisfaction with his or
her job (e.g., “How satisfied are you with the job as a
whole”). All items were answered on a 5-point response
scale (1 ⫽ very satisfied, to 5 ⫽ very dissatisfied).
Results
Analysis Strategy
To accommodate our need to model relations
among variables at multiple levels, data were ana-
lyzed using multilevel structural equation modeling
(ML-SEM) with full maximum likelihood estimation
in Mplus 5.2 (Muthén & Muthén, 1998 –2008). We
used the example Mplus syntax created by Preacher,
199
Zyphur, and Zhang (in press) as a starting point for
developing the syntax that was used to analyze our
multilevel model. In ML-SEM the variability in vari-
ables is decomposed into two latent components, a
within-unit (i.e., individual-level variability) and a
between-unit (i.e., unit-level variability) component
(Lüdtke et al., 2008; Muthén & Asparouhov, 2009).
ML-SEM enables modeling of the relationships
among these variance components within each level
through the specification of measurement and struc-
tural models. At the individual level variables can be
specified as having intercepts (and random slopes)
that vary across units. The random intercepts are
modeled as latent variables at the unit level in which
the respondents within each unit serve as indicators
of the unit level latent construct (Lüdtke et al., 2008;
Muthén & Asparouhov, 2009). In the present study,
we specified no random slopes because we had no
substantive focus on modeling cross-level interactions.
However, random intercepts were specified for the five
EO climate indicators, job stress, and job satisfaction
(see Figure 1). Finally, ML-SEM provides a more
precise estimate of indirect effects in models with
variables at multiple levels of analysis because of the
manner in which variance is decomposed into two
components, thereby avoiding problems of conflated
individual-level and unit-level effects (Preacher et
al., in press; Zhang, Zyphur, & Preacher, 2009).
We followed several steps in conducting our ML-SEM
analyses that generally follow the procedures outlined by
Muthén (1994) and Preacher et al. (in press). As a first
step, we determined the extent of between-unit variance
in all variables by computing Type I intraclass correla-
tion coefficients (ICCs). The ICC(1) reflects the propor-
tion of between-unit variance in a particular variable
(Raudenbush & Bryk, 2002). It is important that vari-
ables have significant between-unit variance to be in-
cluded at the unit level of analysis. In addition, we
computed the rWG(J) statistic (James, Demaree, & Wolf,
1984, 1993) for each unit using a uniform null distri-
bution for the five EO climate indicators and cohesion.
This was done to ensure that a sufficient level of within-
unit agreement was present in the variables for which
we had substantive interest at the unit level. Agreement
was evaluated using LeBreton and Senter’s (2008) re-
vised standards for interpreting interrater agreement es-
timates. For our two unit-level constructs, unit EO cli-
mate and cohesion, we sought a strong level of
agreement to support their inclusion (i.e., median values
greater than or equal to .71; LeBreton & Senter, 2008).
We note that agreement was not calculated for job stress
and job satisfaction because our substantive interest in
the variables was at the individual level.
 200 WALSH, MATTHEWS, TULLER, PARKS, AND MCDONALD

In step two, we estimated a measurement model in Table 3

which all variables within each level were allowed to Results From Analyses To Justify Aggregation
correlate with one another. In the third step, we fit the Median
hypothesized within-unit structural model while al- Construct F p ICC(1) rWG(J)
lowing all constructs at the unit level to freely covary.
Racist behavior 3.53 ⬍.001 .11 .82
In step four, the within-unit and between-unit struc- Race discrimination 2.21 ⬍.001 .05 .92
tural model was specified. Finally, the alternative Age discrimination 1.60 .002 .03 .88
direct paths between unit EO climate and job stress Religious discrimination 1.68 .001 .03 .88
and unit EO climate and job satisfaction were in- Disability discrimination 2.27 ⬍.001 .06 .93
Cohesion 3.06 ⬍.001 .09 .82
cluded to assess the validity of the hypotheses of full Job stress 3.28 ⬍.001 .10 N/A
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

mediation (i.e., Hypotheses 9 and 10). Job satisfaction 3.13 ⬍.001 .09 N/A
This document is copyrighted by the American Psychological Association or one of its allied publishers.

We relied on several model fit indices as a guide

throughout the model building process. In particular,
three measures of model fit were calculated in addition
to the model ␹2: the comparative fit index (CFI), the
root mean square error of approximation (RMSEA),
and the standardized root-mean-square residual
(SRMR). Separate SRMR values are computed for
the within-unit and between-unit models. A nonsig-
nificant ␹2 indicates good model fit; however, be-
cause ␹2 is sensitive to sample size we were con-
cerned primarily with values for the remaining fit
indices in assessing model fit. A CFI value of .95 or
higher, a RMSEA value of .06 or lower, and SRMR
values of .08 or lower indicate good fit of the model
to the data (Hu & Bentler, 1999).
Descriptive Statistics and
Aggregation Analyses
In Table 2, means, standard deviations, internal
consistency reliability information and bivariate cor-
relations for the measures used in the current study
are reported. Review of the means suggest that over-
all respondents perceived a positive EO climate. In
general, relationships among variables at the individ-
ual level were similar to unit-level relationships, and
all correlations were consistent with expectations.
Also, EO climate indicators were significantly corre-
lated with the outcomes at both levels (with the
exception of the relationship between racist behavior
and job satisfaction at the unit level), thus supporting
the first statistical precondition for mediation (Baron
& Kenny, 1986; Mathieu & Taylor, 2006, 2007).
In Table 3 the results from the analyses used to
support aggregation of variables to the unit level are
reported. Significant between-unit variance was ob-
served for all variables with ICCs ranging from .03
(age discrimination, religious discrimination) to .11
(racist behavior). This indicates that anywhere from
3% to 11% of the variance in the constructs is attrib-
utable to unit membership (Raudenbush & Bryk,
Note. The F and p values refer to the F tests and corre-
sponding significance levels from one-way ANOVAs with
unit identifier as the independent variable and the constructs
as the dependent variables.
2002). Further, the median rWG(J) values across units
are presented in Table 3 as well. The median values
range from .82 to .93, indicating that units had strong
agreement on all constructs for which we had sub-
stantive interest at the unit level.
Model Testing
The first step in testing the conceptual multilevel
model presented in Figure 1 was to calculate a mea-
surement model. In the measurement model, the five
EO climate indicators were loaded onto the psycholog-
ical EO climate latent factor in the within-unit model,
and the random intercepts for EO climate indicators
served as indicators for the unit EO climate latent factor
(Lüdtke et al., 2008; Muthén & Asparouhov, 2009).
Race discrimination was set as the marker variable. At
the unit level, job stress and job satisfaction were
represented by their random intercepts. Individual
responses to cohesion were aggregated within each
unit and modeled as an observed variable to avoid
overcomplicating the within-unit model, although we
note that there are limitations associated with aggre-
gation (cf., Lüdtke et al., 2008). At the individual
level, job stress and job satisfaction were modeled as
observed variables.
The measurement model provided good fit to the data
[␹2(30) ⫽ 155.01, p ⬍ .001, CFI ⫽ .97, RMSEA ⫽ .06,
SRMRWithin ⫽ .03, SRMRBetween ⫽ .08], thus we
proceeded to test the within-unit structural model
while letting the constructs freely covary at the unit
level. The model fit indices for the within-unit struc-
tural model were unchanged from the measurement
model (see Table 4); this is not surprising given that
an identical number of parameters were estimated.
 EQUAL OPPORTUNITY CLIMATE
Table 4
Fit Indices for Measurement and Structural Models
Model ␹2 (df)
Measurement model 155.01 (30)
Within-unit structural model 155.01 (30)
Within-unit and between-unit structural model 157.11 (32)
Alternative structural model 155.01 (30)
Note. CFI ⫽ comparative fit index; RMSEA ⫽ root mean square error approximation; SRMRW ⫽ standardized root mean
square residual for the within-unit model; SRMRB ⫽ standardized root mean square residual for the between-unit model.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
Both within-unit and between-unit structural models
were estimated simultaneously in step 3. This model
again provided good fit to the data, [␹2(32) ⫽ 157.01,
p ⬍ .001, CFI ⫽ .97, RMSEA ⫽ .05, SRMRWithin ⫽
.03, SRMRBetween ⫽ .09].
Finally, we examined the fit of an alternative struc-
tural model which included direct paths between unit
EO climate and job stress, and unit EO climate and
job satisfaction. However, the fit of the alternative
model was not significantly better than the fit of the
within-unit and between-unit structural model,
⌬␹2(2) ⫽ 2.10, p ⬎ .05, and neither direct path be-
tween unit EO climate and job stress (␤ ⫽ ⫺.32, p ⬎
.05) or unit EO climate and job satisfaction (␤ ⫽ .04,
p ⬎ .05) was statistically significant. Consequently,
we retained and interpreted the standardized coeffi-
cients for the more parsimonious within-unit and
between-unit structural model. This model is pre-
sented in Figure 2 along with standardized factor
loadings and path estimates, and squared multiple
correlations for endogenous variables are reported in
italics.
In support of Hypothesis 1, psychological EO cli-
mate was positively related to job satisfaction, (␤ ⫽
.27, p ⬍ .001). Hypotheses 2 was also supported as
psychological EO climate was negatively associated
with job stress, (␤ ⫽ ⫺.32, p ⬍ .001), and Hypoth-
eses 3 was supported as job stress negatively pre-
dicted job satisfaction (␤ ⫽ ⫺.33, p ⬍ .001). Like-
wise, the standardized indirect effect of
psychological EO climate on job satisfaction via job
stress was positive and statistically significant (␤ ⫽
.11, p ⬍ .001, 95% CI ⫽ .08, .13), thereby supporting
Hypothesis 4 that job stress partially mediates the
relationship between psychological EO climate and
job satisfaction.3 The model accounted for 10% and
24% of the within-unit (individual-level) variation in
job stress and job satisfaction, respectively.
Unit-level hypotheses were supported as well.
Specifically, a positive relationship between unit EO
climate and cohesion was observed (␤ ⫽ .69, p ⬍
201
p CFI RMSEA SRMRW SRMRB
⬍.001 .97 .06 .03 .08
⬍.001 .97 .06 .03 .08
⬍.001 .97 .05 .03 .09
⬍.001 .97 .06 .03 .08
.001), thus supporting Hypothesis 5. Hypotheses 6
and 7 were supported as cohesion had a significant
positive relation with job satisfaction (␤ ⫽ .64, p ⬍
.001) and a significant negative association with job
stress (␤ ⫽ ⫺.79, p ⬍ .001). Job stress was nega-
tively related to job satisfaction at the unit level as
well (␤ ⫽ ⫺.38, p ⬍ .001), supporting Hypothesis 8.
Furthermore, the standardized total indirect effect of
unit EO climate on job satisfaction was positive and
statistically significant (␤ ⫽ .64, p ⬍ .001, 95% CI ⫽
.38, .90), supporting Hypothesis 9. Hypothesis 10
was also supported as the indirect effect of unit EO
climate on job stress via cohesion was negative and
statistically significant (␤ ⫽ ⫺.54, p ⬍ .001, 95%
CI ⫽ ⫺.80, ⫺.28). These results, in combination
with the lack of direct effects of unit EO climate on
job satisfaction or job stress, support the hypothe-
sized fully mediated relationships between unit EO
climate and job satisfaction and unit EO climate and
job stress. Finally, the model accounted for a sub-
stantial proportion of the between-unit variation in
each endogenous variable, with 47%, 62%, and 92%
of the variance accounted for in cohesion, job stress,
and job satisfaction, respectively. These values may
seem high, but recall that the between-unit variance
in each variable was far lower than the within-unit
variance, thus there was less between-unit variance to
explain.
Analyses To Investigate Potential
Methodological Artifacts Due to Common
Method Variance
One possible methodological concern of field stud-
ies relying on cross-sectional survey data is common
method variance, although Spector (2006) questions
3
Indirect effects were calculated using the MODEL
INDIRECT command in Mplus 5.2 (Muthén & Muthén,
1998 –2008).
 202 WALSH, MATTHEWS, TULLER, PARKS, AND MCDONALD
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
This document is copyrighted by the American Psychological Association or one of its allied publishers.

Figure 2. Results of the empirical multilevel model. RB ⫽ racist behavior; RD ⫽ race

discrimination; AD ⫽ age discrimination; RLD ⫽ religious discrimination; and DD ⫽
disability discrimination. Squared multiple correlations for endogenous variables are reported
in italics. ⴱ p ⬍ .05. ⴱⴱⴱ p ⬍ .001.

how widespread the issue really is in survey research.
Nonetheless, our measures were self-reported from a
single point in time which raises the possibility that
observed relationships may have been spuriously in-
flated. We note that common method variance should
only be considered a serious issue if there appears to
be a systematic and pervasive inflation of observed
relationships (James, Gent, Hater, & Corey, 1979). In
our study, the range of individual-level correlations
(r ⫽ ⫺.22 to .69), absence of multicollinearity
(Kline, 1998), and absence of nonintuitive relation-
ships leads us to believe that common method vari-
ance is not a significant concern.
As an additional empirical check to assess issues
of common method variance in our data, using the
full sample (N ⫽ 1,394) a principal-axis factor anal-
ysis (obliman rotation, ⌬ ⫽ 0) with Kaiser normal-
ization was conducted on the five EO climate indi-
cators and items for the remaining constructs. This
analysis resulted in a four-factor solution explaining
70.7% of the variance. Eigenvalues ranged from 1.32
to 7.14 and each factor uniquely represented one of
the four constructs included in the model. As noted
by Podsakoff, MacKenzie, Lee, and Podsakoff
(2003), although this method has its limitations, the
fact that four distinct factors, and not a single factor,
accounted for the majority of variance in the solution
suggests that common method variance is not a sig-
nificant issue with our data.
Discussion
The goal of this study was to extend EO climate
research by testing a theoretically grounded concep-
tual multilevel model. We anticipated that EO cli-
mate would operate at the individual and unit levels
of analysis, with psychological EO climate exerting
direct effects on job satisfaction and job stress and
unit EO climate exerting cross-level indirect effects
on job satisfaction and job stress via cohesion. To our
knowledge, no research has examined EO climate as
a shared property of the unit and little attention has
been given to mechanisms linking EO climate to
 EQUAL OPPORTUNITY CLIMATE
various work-related attitudes and perceptions. Thus,
the current research provides several contributions to
the EO climate literature.
As predicted, psychological EO climate had a di-
rect positive effect on employees’ job satisfaction,
which replicates findings from several previous stud-
ies (e.g., Estrada & Harbke, 2008; McIntyre et al.,
2002). Psychological EO climate also had a direct
negative influence on self-reported job stress. Of the
research to date, this is the first study we are aware of
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
to find evidence of direct effects of psychological EO
This document is copyrighted by the American Psychological Association or one of its allied publishers.
climate on job stress, although this is consistent with
a broader literature linking fairness perceptions to
stress (e.g., Judge & Colquitt, 2004) and findings on
effects of other psychological climates (e.g., Kessler
et al., 2008). Furthermore, job stress was a negative
predictor of job satisfaction, which is consistent with
previous work demonstrating significant, negative as-
sociations between job stress and work attitudes
(Parasuraman & Alutto, 1984; Summers et al., 1995).
Finally, job stress partially mediated the relationship
between psychological EO climate and job satisfac-
tion, implying that more positive EO climate percep-
tions are associated with lower job stress and greater
job satisfaction.
We expected that unit members would report sim-
ilar perceptions of the EO climate within their work-
place, based on the rationale that individuals working
in the same location are exposed to similar EO pol-
icies, practices, procedures, and behavioral norms
(Schneider & Reichers, 1983). A high level of agree-
ment was observed with respect to within-unit ratings
of EO climate, indicating that EO climate functions
not only at the individual level as a psychological
climate but also as a shared property of the unit
(Kozlowski & Klein, 2000). Effects of unit EO cli-
mate were hypothesized given the theoretical and
empirical justification for studying unit EO climate.
Consistent with our predictions, unit EO climate
was positively associated with cohesion. This link is
especially important within the military due to the
impact of cohesion on mission effectiveness, reten-
tion, and well-being (Oliver et al., 1999). This link is
also consistent with fairness heuristic theory which
suggests that relational constructs such as cohesion
should be strongly associated with general fairness
perceptions (Lind, 2001). Furthermore, cohesion was
positively related to job satisfaction and negatively
related to job stress. These cross-level direct effects
are consistent with prior research and theory on the
influence of cohesion (e.g., Oliver et al., 1995; Parker
& DeCotiis, 1983; Wech et al., 1998). In addition,
fully mediated relationships were found between unit
203
EO climate and job satisfaction and unit EO climate
and job stress. This evidence suggests that the pro-
cesses involved in transmitting effects of EO climate
differ depending on the level at which EO climate is
operationalized. Although psychological EO climate
demonstrated direct effects on job stress and job
strain (and indirect effects on job satisfaction via
stress), unit EO climate exerted cross-level influences
on the same variables through cohesion.
These findings point to the need for researchers to
continue studying effects of psychological and higher
level climates not only in future EO climate research,
but also in research on other climate constructs. The
recent meta-analysis by Parker et al. (2003) on effects
of psychological climate and review by Kuenzi and
Schminke (2009) on organizational climate show that
climate constructs studied at various levels have
meaningful influences on a host of work-related atti-
tudes and perceptions. Based on this work and find-
ings from the current research, there is sufficient
evidence supporting the need for continued investi-
gation into effects of psychological, unit, and orga-
nizational climates.
Collectively, the results bolster the importance of
developing fair and equitable work environments as
well as developing a comprehensive understanding of
organizational fairness, one which incorporates EO
fairness perceptions. It is somewhat surprising that
relatively little attention has been paid to fairness
perceptions surrounding EO, particularly in light of
findings from the present study. Recall that EO cli-
mate is not directly associated with personal experi-
ences of discrimination or harassment, rather EO
climate concerns how likely it is that such forms of
unfair treatment are thought to occur. Nonetheless,
these individual and shared perceptions have a note-
worthy influence as seen here and in previous re-
search (e.g., Estrada & Harbke, 2008; Estrada et al.,
2007; McIntyre et al., 2002). Similar to Lind (2001),
we contend that creating and maintaining fair work
practices and environments will likely be met with
positive outcomes including those investigated in the
present study and beyond. We also suggest that re-
searchers continue to study EO climate from an or-
ganizational justice and fairness perspective because
underlying EO climate are concerns of fair and just
treatment. We feel that the greatest gains in our
understanding of EO climate will only be made if
working from well-developed theory such as that
offered in the organizational justice and fairness lit-
eratures.
 204 WALSH, MATTHEWS, TULLER, PARKS, AND MCDONALD
Limitations and Future Directions
Findings from this study continue to shed light on
the influence of EO climate. However, the current
study is limited in several ways and future research is
needed to address these limitations. First, we tested
our model within a sample of active duty military per-
sonnel which is consistent with previous research on the
construct (Dansby & Landis, 1991; Edwards, 2001;
Truhon, 2008). However, the generalizability of the
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
current findings to individuals in other organizations
This document is copyrighted by the American Psychological Association or one of its allied publishers.
and industries is unknown. As such, there is a need to
assess the influence of EO climate outside the context of
the military. There is ample evidence demonstrating the
significant influence of fairness perceptions across a
variety of settings (e.g., Colquitt et al., 2001). It serves
to reason, then, that EO climate likely plays an impor-
tant role in influencing work attitudes in organizational
settings beyond the military.
Additionally, our sample consisted of military per-
sonnel that were not currently deployed due to concerns
about differences between deployed and nondeployed
personnel (e.g., Adler et al., 2005). However, given the
current global presence of the United States military,
either in peace keeping roles or combat missions, vali-
dating the model among deployed personnel constitutes
an important next step. We may expect divergent find-
ings between deployed and nondeployed personnel for
one or more relationships in our model. For example,
there may be an even stronger negative relationship
between cohesion and job stress among deployed per-
sonnel. Deployed personnel may experience decreased
social support from sources outside the military, due to
diminished time with common support systems such as
family and friends, thus making one’s unit the primary
means of social support. Ultimately, the validation of
this model in a deployed sample could enhance our
understanding of EO climate and further demonstrate its
importance in terms of mission effectiveness.
Further, the referent for the cohesion measure was
inconsistent with the level at which the data were
aggregated. That is, cohesion items referred to one’s
work group whereas the data were aggregated to the
level of the military unit. This inconsistency slightly
clouds the interpretation of the results from this study
as we cannot explicitly state that our measure of
cohesion accurately reflects unit cohesion. However,
we did not have access to any identifiers at the level
of the work group, only at the unit level, thus data
could not be aggregated to any level other than the
military unit. We felt that the advantages of aggre-
gating cohesion (e.g., the ability to conceptualize and
model effects of EO climate at two levels) out-
weighed the disadvantages associated with the incon-
sistent referent. What’s more, our analyses to provide
empirical justification for aggregating cohesion indi-
cated levels of agreement that are consistent with
recommended levels. While we feel that the refer-
ent—level of analysis inconsistency is a limitation of
the study, we do not feel that this limitation is so
problematic as to invalidate the conclusions drawn
from the findings. However, future multilevel EO
climate research would benefit from having referents
that are consistent with the unit of analysis.
Although our conceptual model implies a logical
ordering of relations among variables based on theory,
due to the cross-sectional study design we caution read-
ers against inferring causality in the relationships
among constructs. A related limitation of cross-
sectional research is the potential concern of common
method bias. As described earlier, to address this con-
cern we conducted a principal-axis factor analysis of the
items used to assess each of our constructs, with results
suggesting a four-factor model in which each of the
resulting factors represented a construct in the model.
These results minimize our concerns to some extent;
however, it is still important to consider our findings in
light of such limitations associated with self-report,
cross-sectional survey research designs.
In addition to the aforementioned limitations
which should be addressed in future research, there
are several other avenues for future EO climate work
that can build on this study. In the current study,
support was found for job stress as a partial mediator
of the relationship between psychological EO climate
and job satisfaction, and cohesion was a key mediator
at the unit level. Future research should attempt to
replicate these observations on additional work-
related outcomes. Likewise, it will be important for
future EO climate research to consider additional
unit-level and individual-level variables that mediate
effects of EO climate. For example, Raver and Gel-
fand (2005) found support for team conflict and team
cohesion as mediators of the relationship between
ambient sexist hostility and team financial perfor-
mance. Although Raver and Gelfand (2005) focused
on sexual harassment, EO climate shares a similar
theme pertaining to respectful treatment in the work-
place, so team or unit conflict may be one additional
mediator of effects of EO climate on work-related
attitudes and perceptions.
Furthermore, recent work suggests that it is not
only important to consider climate level but also the
strength of the climate when conducting climate re-
search, and that relationships between climate and
outcomes are generally greater within strong climates
 EQUAL OPPORTUNITY CLIMATE
(i.e., low within-unit variability; e.g., Colquitt et al.,
2002; Schneider, Salvaggio, & Subirats, 2002). Re-
lated to the current study, the relationship between
unit EO climate and cohesion (and other constructs
related to team processes) may be stronger in units
with a strong EO climate. In sum, we envision a
number of possibilities for future EO climate re-
search, and we urge researchers to pursue future
investigations into the influence of psychological and
unit EO climate on work-related attitudes and behav-
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
ior.
This document is copyrighted by the American Psychological Association or one of its allied publishers.
References
Adams, J. S. (1965). Inequity in social exchange. In L.
Berkowitz (Ed.), Advances in experimental social psy-
chology (Vol. 2, pp. 267–299). New York: Academic
Press.
Adler, A. B., Huffman, A. H., Bliese, P. D., & Castro, C. A.
(2005). The impact of deployment length and experience
on the well-being of male and female soldiers. Journal of
Occupational Health Psychology, 10, 121–137.
Asparouhov, T., & Muthén, B. (2006). Constructing covari-
ates in multilevel regression. Mplus Web Notes: No. 11.
Available from: Http://www.statmodel.com
Baron, R. M., & Kenny, D. A. (1986). The moderator-
mediator variable distinction in social psychological re-
search: Conceptual, strategic, and statistical consider-
ations. Journal of Personality and Social Psychology, 51,
1173–1182.
Bies, R. J., & Moag, J. S. (1986). Interactional justice:
Communication criteria of fairness. In R. J. Lewicki,
B. H. Sheppard, & M. Bazerman (Eds.), Research on
negotiation in organizations (pp. 43–55). Greenwich,
CT: JAI Press.
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 Psy-
chology, 83, 234 –246.
Colquitt, J. A., Conlon, D. E., Wesson, M. J., Porter, C., &
Ng, K. Y. (2001). Justice at the millennium: A meta-
analytic review of 25 years of organizational justice
research. Journal of Applied Psychology, 86, 425– 445.
Colquitt, J. A., Noe, R. A., & Jackson, C. L. (2002). Justice
in teams: Antecedents and consequences of procedural
justice climate. Personnel Psychology, 55, 83–109.
Cooper, C. L., & Marshall, J. (1976). Occupational sources
of stress: A review of the literature relating to coronary
heart disease and mental ill health. Journal of Occupa-
tional Psychology, 49, 11–28.
Cox, T. (1993). Cultural diversity in organizations. San
Francisco: Berrett-Koehler.
Dansby, M. R., & Landis, D. (1991). Measuring equal
opportunity in the military environment. International
Journal of Intercultural Relations, 15, 389 – 405.
Dansby, M. R., & Landis, D. (1998). Race, gender, and
representation index as predictors of an equal opportunity
climate in military organizations. Military Psychology,
10, 87–105.
Defense Equal Opportunity Management Institute
(DEOMI) Program Course Catalog. (2007). Retrieved
205
from http://www.deomi.org/Education& Training/
documents/CourseCatalog_2007_v2.pdf
Defense Manpower Data Center (DMDC). (2007). Active
duty demographic profile. Retrieved from http://
www.deomi
.org/EOEEOResources/DemographicReports.cfm
Edwards, J. E. (2001). Opportunities for assessing military
EO: A researcher’s perspective on identifying an integra-
tive program-evaluation strategy. In M. R. Dansby, J. B.
Stewart, & S. C. Webb (Eds.), Managing diversity in the
military: Research perspectives from the Defense Equal
Opportunity Management Institute (pp. 163–177). New
Brunswick, NJ: Transaction Publishers.
Estrada, A. X., & Harbke, C. R. (2008). Gender and ethnic
differences in perceptions of equal opportunity climate
and job outcomes of US Army Reserve component per-
sonnel. International Journal of Intercultural Relations,
32, 466 – 478.
Estrada, A. X., Stetz, M. C., & Harbke, C. R. (2007).
Further examination and refinement of the psychometric
properties of the MEOCS with data from reserve com-
ponent personnel. International Journal of Intercultural
Relations, 31, 137–161.
Executive Order No. 9981, 13 C. F. R. 4313. (1948).
Folger, R. (1987). Reformulating the preconditions of re-
sentment: A referent cognitions model. In J. C. Masters
& W. P. Smith (Eds.), Social Comparison, justice, and
relative deprivation: Theoretical, empirical, and policy
perspectives (pp. 183–215). Hillsdale, NJ: Erlbaum.
Folger, R. (1993). Reactions to mistreatment at work. In
J. K. Murnighan (Ed.), Social psychology in organiza-
tions: Advances in theory and research (pp. 161–183).
Englewood Cliffs, NJ: Prentice Hall.
Folger, R., & Cropanzano, R. (2001). Fairness theory: Jus-
tice as accountability. In J. Greenberg & R. Cropanzano
(Eds.), Advances in organizational justice (pp. 1–55).
Stanford, CA: Stanford University Press.
Gilson, L. L., Mathieu, J. E., Shalley, C. E., & Ruddy, T. M.
(2005). Creativity and standardization: Complementary
or conflicting drivers of team effectiveness. Academy of
Management Journal, 48, 521–531.
Glomb, T. M., & Liao, H. (2003). Interpersonal aggression
in work groups: Social influence, reciprocal, and individ-
ual effects. Academy of Management Journal, 46, 486 –
496.
Griffith, J., & Vaitkus, M. (1999). Relating cohesion to
stress, strain, disintegration, and performance: An orga-
nizing framework. Military Psychology, 11, 27–55.
Gully, S. M., Devine, D. J., & Whitney, D. J. (1995). A
meta-analysis of cohesion and performance: Effects of
level of analysis and task interdependence. Small Group
Research, 26, 497–520.
Hicks-Clarke, D., & Iles, P. (2000). Climate for diversity
and its effects on career and organizational attitudes and
perceptions. Personnel Review, 29, 324 –345.
Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit
indexes in covariance structure analysis: Conventional
criteria versus new alternatives. Structural Equation
Modeling, 6, 1–55.
James, L., Gent, M., Hater, J., & Corey, K. (1979). Corre-
lates of psychology influence: An illustration of the psy-
chological climate approach to work environment. Per-
sonnel Psychology, 32, 563–588.
James, L. R., Demaree, R. G., & Wolf, G. (1984). Estimat-
 206 WALSH, MATTHEWS, TULLER, PARKS, AND MCDONALD
ing within-group interrater reliability with and without
response bias. Journal of Applied Psychology, 69, 85–98.
James, L. R., Demaree, R. G., & Wolf, G. (1993). rWG: An
assessment of within-group interrater agreement. Journal
of Applied Psychology, 78, 306 –309.
James, L. R., & Jones, A. P. (1974). Organizational climate:
A review of theory and research. Psychological Bulletin,
81, 1096 –1112.
Judge, T. A., & Colquitt, J. A. (2004). Organizational jus-
tice and stress: The mediating role of work-family con-
flict. Journal of Applied Psychology, 89, 395– 404.
Kessler, S. R., Spector, P. E., Chang, C., & Parr, A. D.
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
(2008). Organizational violence and aggression: Devel-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
opment of the three-factor Violence Climate Survey.
Work & Stress, 22, 108 –124.
Kline, R. B. (1998). Principles and practice of structural
equation modeling. New York: Guilford Press.
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– 81.
Kozlowski, S. W. J., & Klein, K. J. (2000). A multilevel
approach to theory and research in organizations: Contex-
tual, temporal, and emergent processes. In K. Klein and S.
Kozlowski (Eds.), Multilevel theory, research, and methods
in organizations: Foundations, extensions, and new direc-
tions (pp. 3–90). San Francisco: Jossey-Bass.
Kuenzi, M., & Schminke, M. (2009). Assembling fragments
into a lens: A review, critique, and proposed research
agenda for the organizational work climate literature.
Journal of Management, 35, 634 –717.
Landis, D., Dansby, M. R., & Faley, R. H. (1993). The
Military Equal Opportunity Climate Survey: An example
of surveying in organizations. In P. Rosenfeld, J. E.
Edwards, & M. D. Thomas (Eds.), Improving organiza-
tional surveys: New directions, methods, and applica-
tions (pp. 210 –239). Newbury Park, CA: Sage.
Landis, D., Fisher, G., & Dansby, M. R. (1988). Construc-
tion and preliminary validation of an equal opportunity
climate assessment instrument. In F. E. McIntire (Ed.),
Proceedings of Psychology in the DoD Symposium
(Tech. Rep. No. 88-1, pp. 487– 491). Colorado Springs,
CO: U. S. Air Force Academy.
Lazarus, R. S. (1981). Little hassles can be hazardous to
health. Psychology Today, 15, 58 – 62.
LeBreton, J. M., & Senter, J. L. (2008). Answers to 20
questions about interrater reliability and interrater agree-
ment. Organizational Research Methods, 11, 815– 852.
Liao, H., & Rupp, D. E. (2005). The impact of justice
climate and justice orientation on work outcomes: A
cross-level multifoci framework. Journal of Applied Psy-
chology, 90, 242–256.
Lind, E. A. (2001). Fairness heuristic theory: Justice judg-
ments as pivotal cognitions in organizational relations. In
J. Greenberg & R. Cropanzano (Eds.), Advances in or-
ganizational justice (pp. 56 – 88). Stanford, CA: Stanford
University Press.
Lind, E. A., & Tyler, T. R. (1988). The social psychology of
procedural justice. New York: Plenum Press.
Lüdtke, O., Marsh, H. W., Robitzsch, A., Trautwein, U.,
Asparouhov, T., & Muthén, B. (2008). The multilevel
latent covariate model: A new, more reliable approach to
group-level effects in contextual studies. Psychological
Methods, 13, 203–229.
Marquis, J. P., Lim, N., Scott, L. M., Harrell, M. C., &
Kavanagh, J. (2008). Managing diversity in corporate
America. An exploratory analysis. Retrieved from http://
www.rand.org/pubs/occasional_papers/2007/RAND_
OP206.pdf
Mathieu, J. E., & Taylor, S. R. (2006). Clarifying conditions
and decision points for mediational type inferences in
organizational behavior. Journal of Organizational Be-
havior, 27, 1031–1056.
Mathieu, J. E., & Taylor, S. R. (2007). A framework for
testing meso-mediational relationships in organizational
behavior. Journal of Organizational Behavior, 28, 141–
172.
McIntyre, R. M., Bartle, S. A., Landis, D., & Dansby, M. R.
(2002). The effects of equal opportunity fairness attitudes
on job satisfaction, organizational commitment, and per-
ceived work group efficacy. Military Psychology, 14,
299 –319.
Mor Barak, M. E., Cherin, D. A., & Beckman, S. (1998).
Organizational and personal dimensions of diversity cli-
mate. Journal of Applied Behavioral Science, 34, 82–
105.
Muthén, B., & Asparouhov, T. (2009). Beyond multilevel
regression modeling: Multilevel analysis in a general
latent variable framework. To appear in J. Hox & J. K.
Roberts (Eds.), The handbook of advanced multilevel
analysis. Taylor and Francis.
Muthén, B. O. (1994). Multilevel covariance structure anal-
ysis. Sociological Methods and Research, 22, 376 –398.
Muthén, L. K., & Muthén, B. O. (1998 –2008). Mplus user’s
guide: Statistical analysis with latent variables (5th ed.).
Los Angeles: Muthén & Muthén.
Naumann, S. E., & Bennett, N. (2000). A case for proce-
dural justice climate: Development and test of a multi-
level model. Academy of Management Journal, 43, 881–
889.
Newell, C. E., Rosenfeld, P., & Culbertson, A. L. (1995).
Sexual harassment experiences and equal opportunity
perceptions of Navy women. Sex Roles, 32, 159 –168.
Oliver, L. W., Harman, J., Hoover, E., Hayes, S. M., &
Pandhi, N. A. (1999). A quantitative integration of the
military cohesion literature. Military Psychology, 11,
57– 83.
Parasuraman, S., & Alutto, J. A. (1984). Sources and out-
comes of stress in organizational settings: Toward the
development of a structural model. Academy of Manage-
ment Journal, 27, 330 –350.
Parker, C. P., Baltes, B. B., Young, S. A., Huff, J. W.,
Altmann, R. A., LaCost, H. A., & Roberts, J. E. (2003).
Relationships between psychological climate perceptions
and work outcomes: A meta-analytic review. Journal of
Organizational Behavior, 24, 389 – 416.
Parker, D. F., & DeCotiis, T. A. (1983). Organizational
determinants of job stress. Organizational Behavior and
Human Performance, 32, 160 –177.
Podsakoff, P. M., MacKenzie, S. B., Lee, J. Y., & Podsa-
koff, N. P. (2003). Common method biases in behavioral
research: A critical review of the literature and recom-
mended remedies. Journal of Applied Psychology, 88,
879 –903.
Preacher, K., Zyphur, M. J., & Zhang, Z. (in press). A
general multilevel SEM framework for assessing multi-
level mediation. Psychological Methods.
Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical
 EQUAL OPPORTUNITY CLIMATE
linear models: Applications and data analysis methods
(2nd ed.). Thousand Oaks, CA: Sage.
Raver, J. L., & Gelfand, M. J. (2005). Beyond the individual
victim: Linking sexual harassment, team processes, and
team performance. Academy of Management Journal, 48,
387– 400.
Rosenfeld, P., Newell, C. E., & Le, S. (1998). Equal op-
portunity climate of women and minorities in the Navy:
Results from the Navy equal opportunity/sexual harass-
ment (NEOSH) survey. Military Psychology, 69 – 85.
Rousseau, D. M. (1988). The construction of climate in
organizational research. In C. L. Cooper & I. Robertson
This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.
(Eds.), International review of industrial and organiza-
This document is copyrighted by the American Psychological Association or one of its allied publishers.
tional psychology (pp. 137–158). London: Wiley.
Schneider, B., & Reichers, A. E. (1983). On the etiology of
climates. Personnel Psychology, 36, 19 –39.
Schneider, B., Salvaggio, A. N., & Subirats, M. (2002).
Climate strength: A new direction for climate research.
Journal of Applied Psychology, 87, 220 –229.
Schneider, B., White, S. S., & Paul, M. C. (1998). Linking
service climate and customer perceptions of service qual-
ity: Test of a causal model. Journal of Applied Psychol-
ogy, 83, 150 –163.
Selye, H. (1976). The stress of life. New York: McGraw-
Hill.
Spector, P. E. (2006). Method variance in organizational
research: Truth or urban legend? Organizational Re-
search Methods, 9, 221–232.
Spector, P. E., Coulter, M. L., Stockwell, H. G., & Matx,
M. W. (2007). Relationships of workplace physical vio-
lence and verbal aggression with perceived safety, per-
ceived violence climate, and strains in a healthcare set-
ting. Work & Stress, 21, 117–130.
Spector, P. E., & Jex, S. M. (1998). Development of four
self-report measures of job stressors and strain: Interper-
sonal conflict at work scale, organizational constraints scale,
quantitative workload inventory, and physical symptoms
207
inventory. Journal of Occupational Health Psychology,
3, 356 –367.
Summers, T. P., DeCotiis, T. A., & DeNisi, A. S. (1995). A
field study of some antecedents and consequences of felt
job stress. In R. Crandall & R. L. Perrewe (Eds.), Occu-
pational Stress: A Handbook (pp. 113–128). Washing-
ton, DC: Taylor and Francis.
Truhon, S. A. (2003, November). DEOCS: A new and
improved MEOCS. Paper presented at the 45th Annual
Conference of the International Military Testing Organi-
zation, Pensacola, FL.
Truhon, S. A. (2008). Equal opportunity climate in the
United States military: Are differences in the eye of the
beholder? European Journal of Work and Organizational
Psychology, 17, 153–169.
Truhon, S. A., & Parks, K. M. (2007, October). Norming the
DEOCS organizational climate survey. Paper presented
at the 49th International Military Testing Association
Conference, Gold Coast, Queensland, Australia.
Tyson, A. S. (2008, September 5). Soldiers’ suicide rate on
pace to set record. Washington Post, p. A02.
Wech, B. A., Mossholder, K. W., Steel, R. P., & Bennett, N.
(1998). Does work group cohesiveness affect individu-
als’ performance and organizational commitment? A
cross-level examination. Small Group Research, 29,
472– 494.
Zhang, Z., Zyphur, M. J., & Preacher, K. J. (2009). Testing
multilevel mediation using hierarchical linear models:
Problems and solutions. Organizational Research Meth-
ods, 12, 695–719.
Zohar, D. (1980). Safety climate in industrial organizations:
Theoretical and applied implications. Journal of Applied
Psychology, 65, 96 –102.
Received March 24, 2009
Revision received October 3, 2009
Accepted November 29, 2009 y