Accident Analysis and Prevention 99 (2017) 77–89

Contents lists available at ScienceDirect

Accident Analysis and Prevention

journal homepage: www.elsevier.com/locate/aap

Disentangling the roles of safety climate and safety culture:

Multi-level effects on the relationship between supervisor
enforcement and safety compliance
Laura Petitta a,∗ , Tahira M. Probst b , Claudio Barbaranelli a , Valerio Ghezzi a
a
Sapienza – University of Rome, Italy
b
Washington State University Vancouver, USA

a r t i c l e i n f o a b s t r a c t

Article history: Despite increasing attention to contextual effects on the relationship between supervisor enforcement
Received 22 July 2015 and employee safety compliance, no study has yet explored the conjoint influence exerted simulta-
Received in revised form 1 June 2016 neously by organizational safety climate and safety culture. The present study seeks to address this
Accepted 11 November 2016
literature shortcoming. We first begin by briefly discussing the theoretical distinctions between safety
climate and culture and the rationale for examining these together. Next, using survey data collected
Keywords:
from 1342 employees in 32 Italian organizations, we found that employee-level supervisor enforce-
Safety
ment, organizational-level safety climate, and autocratic, bureaucratic, and technocratic safety culture
Supervisor
Compliance
dimensions all predicted individual-level safety compliance behaviors. However, the cross-level moder-
Climate ating effect of safety climate was bounded by certain safety culture dimensions, such that safety climate
Culture moderated the supervisor enforcement-compliance relationship only under the clan-patronage culture
dimension. Additionally, the autocratic and bureaucratic culture dimensions attenuated the relation-
ship between supervisor enforcement and compliance. Finally, when testing the effects of technocratic
safety culture and cooperative safety culture, neither safety culture nor climate moderated the relation-
ship between supervisor enforcement and safety compliance. The results suggest a complex relationship
between organizational safety culture and safety climate, indicating that organizations with particular
safety cultures may be more likely to develop more (or less) positive safety climates. Moreover, employee
safety compliance is a function of supervisor safety leadership, as well as the safety climate and safety
culture dimensions prevalent within the organization.
© 2016 Elsevier Ltd. All rights reserved.

1. Introduction result, the European Occupational Safety & Health Administration

Lack of workplace safety can adversely affect workers’ lives,
represents a social cost through an increase of health and social
security demands, and causes disruptions to production and/or
negative publicity for businesses and organisations (European
Commission, 2011). A 2006 report from the International Labor
Organization (ILO) estimates that the total costs of work-related
accidents and ill-health equate to approximately 4% of the worlds
Gross Domestic Product (ILO, 2006). Within Italy alone, more
than 500,000 workers are injured annually (Istituto Nazionale
per l’Assicurazione contro gli Infortuni sul Lavoro-INAIL [National
Institute for Insurance against Accidents at Work], 2014). As a
∗ Corresponding author at: Department of Psychology, Sapienza University of
Rome, Via dei Marsi, 78, 00185, Rome, Italy.
E-mail address: laura.petitta@uniroma1.it (L. Petitta).
(EU-OSHA, 2002) is concerned about these costs for Europe and
has prioritised the need to develop knowledge of the economic
and social costs arising from such safety failures and how best
to prevent such failures. Similarly, according to the US Occupa-
tional Safety & Health Administration (OSHA), the direct costs of
work-related injuries and illnesses include workers’ compensation
payments, medical expenses, and costs for legal services, whereas
indirect costs include training replacement employees, accident
investigation and implementation of corrective measures, lost pro-
ductivity, repairs of damaged equipment and property, and costs
associated with lower employee morale and absenteeism.
In addition to relying on prevention through design, advanced
ergonomics and personal protective equipment, a growing body
of literature conducted largely by occupational and organizational
researchers has increasingly focused on the influence of supervisor-
and organizational-level psychosocial variables on employee safety

http://dx.doi.org/10.1016/j.aap.2016.11.012
0001-4575/© 2016 Elsevier Ltd. All rights reserved.
78 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89
performance and outcomes (e.g., Christian et al., 2009; Clarke,
2006). Specifically, a great deal of attention has been paid to the
contextual effects of organizational safety culture and safety cli-
mate. Unfortunately, despite the advances that have occurred since
Zohar’s (1980) seminal article on safety climate, there has been
pervasive and persistent conflation of the terms safety culture and
safety climate which has hindered our understanding of how these
contextual variables might jointly influence employee safety out-
comes.
Specifically, there has been a lack of clear theoretical and
measurement distinctions between safety climate and culture con-
structs among scientists and practitioners alike. Indeed, as noted by
other scholars, the two terms are often used interchangeably (Cox
and Flin, 1998; Guldenmund, 2007; Mearns and Flin, 1999; Yule,
2003). Therefore, the purpose of our study is to briefly describe the
theoretical differences between safety climate and safety culture
and empirically test their relationships with and purported effects
on employee safety compliance. We use a multilevel modelling
approach to test our model by investigating the extent to which the
relationship between supervisor enforcement and employee safety
compliance is moderated by organizational-level safety climate and
safety culture. By examining the conjoint influence exerted simul-
taneously by organizational climate and culture, we hope to gain a
better understanding of the relationship between these two impor-
tant constructs and identify subsequent implications for science
and practice.
Fig. 1 illustrates the empirical model and hypotheses to be tested
positing that at the individual-level employee safety compliance
is influenced by supervisor enforcement. However, we also pro-
pose that this relationship is moderated by both organizational
safety climate and safety culture (which are also expected to exert
direct main effects on compliance). In positing this model, we do
not intend to imply that these are the only individual and contex-
tual variables that influence safety behavior. However, the model
does seek to address a literature shortcoming by investigating the
simultaneous and multilevel moderating roles of safety climate and
culture on the supervisor enforcement-compliance relationship.
Researchers have extensively discussed and empirically inves-
tigated the concepts of culture and climate (cf., Ashkanasy et al.,
2010; Cooper et al., 2001; House et al., 2004); yet, this study rep-
resents the first attempt to empirically test their conjoint and
simultaneous contributions to the understanding of employee
safety. In doing so, we hope to provide a complementary approach
to ongoing research, which has tended to alternatively focus on
either safety culture or safety climate. Moreover, our empirical
results may have important theoretical and practical implications,
given the emphasis that is placed on developing effective means of
improving employee compliance and reducing workplace injuries.
We begin our review of this literature by examining the impor-
tance of focusing on the individual-level relationship between
supervisor enforcement and safety compliance. We next briefly
define and delineate the constructs of organizational safety climate
and safety culture, and consider how these organizational-level
variables might influence the individual-level relationship between
supervisor enforcement and safety compliance.
2. The relationship between supervisor enforcement and
safety compliance
Although workplace injuries and accidents are the ultimate
endpoints in the causal chain typically examined by workplace
safety researchers, such variables tend to be (fortunately) low
base rate phenomena. However, contemporary models of work-
place safety (e.g., Brondino et al., 2012) highlight the important
role of safety performance as precursors to these important work-
place outcomes. While Griffin and Neal (2000) argued that safety
performance includes both safety compliance and safety participa-
tion, safety participation refers to discretionary behaviors such as
helping co-workers improve safety in the workplace (i.e, safety cit-
izenship behaviors), whereas safety compliance includes the core
compliance-related activities to maintain safety, such as wearing
personal protective equipment (Neal et al., 2000). Hence, safety
compliance can be defined as the extent to which employees adhere
to safety procedures and carry out their work tasks according to
prescribed safety rules, regulations, and policies (Neal et al., 2000).
A recent meta-analysis by Nahrgang et al. (2011) found that safety
compliance was more strongly correlated with workplace accidents
and injuries (corrected r = −0.20), as well as other adverse safety-
related events (corrected r = −0.49), compared to the relationship
between safety participation and these two outcomes (corrected
r = −0.08 and corrected r = −0.32, respectively). Therefore, we chose
to focus on safety compliance as our dependent variable of interest,
rather than safety participation.
There is considerable research evidence indicating the impor-
tance of supervisor safety-specific behaviors in predicting
employee safety compliance and safety-related outcomes (see
Hofmann and Morgeson, 2004 for a review). For example, accord-
ing to Zohar (2003), enacted organizational safety policies refer to
the actual implementation and execution of safety procedures
and practices among supervisors and employees. In other words,
actions speak louder than words when determining safety’s true
prominence within an organization. In keeping with this idea,
Probst and Brubaker (2001) proposed that extrinsic safety moti-
vation (i.e., supervisor enforcement) would be related to employee
safety compliance at work. Extrinsic safety motivation involves the
perceptions of supervisor enforcement of safety policies, including
the extent to which supervisors provide praise for safety compli-
ance and punish for non-compliance. They found that employees
who had low extrinsic safety motivation (i.e., supervisors who
failed to enforce safety policies) had lower levels of safety compli-
ance and were more likely to experience injuries and accidents at
work. Thus, this aspect of enacted safety policy was shown to influ-
ence safety outcomes. More recently, Fugas et al. (2012) focused
on a combination of cognitive and social mechanisms, finding
that supervisors’ injunctive safety norms and perceived behav-
ioral control predict workers’ compliance safety behaviors. Indeed,
a subsequent meta-analysis by Clarke (2013) found that transac-
tional safety leadership (i.e., a focus on supervisor enforcement)
was more predictive of safety compliance than transformational
leadership (which tended to be more related to discretionary safety
behaviors).
In keeping with these findings, we expect that:
Hypothesis 1 (replication): Greater supervisor safety enforce-
ment will be related to higher levels of employee safety compliance.
In addition to proposing a direct relationship between super-
visor enforcement and safety compliance, we also expect that the
organizational safety culture and organizational safety climate will
moderate this relationship. Before developing specific hypotheses
regarding these relationships, it is important to briefly clarify the
theoretical and practical distinctions between safety climate and
safety culture, particularly as these terms are often used inter-
changeably in the literature and in practice. 3. Safety Culture and
Safety Climate
In the realm of safety studies, despite a general agreement
among scholars that climate and culture are different concepts,
multiple often-overlapping definitions of the two constructs have
been put forth in the literature. As a result, safety climate and safety
culture are often confused in theory and practice despite having
distinct etymology (Cox and Flin, 1998). Furthermore, according to
Yule (2003), a review of the definitions of safety culture and safety
 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89
Organizational
Safety
Climate
Organizational-level
Individual-level
Supervisor
Enforcement
Fig. 1. Empirical model tested in the current study.
climate leads to the conclusion that the vast majority of research
on safety culture actually reflects safety climate instead.
The following sections will briefly review some theoretical pro-
posals of safety climateand safety culture and develop the empirical
model including safety climate and culture to be tested in the
present study.
2.1. Safety climate
The term safety climate was first defined by Zohar (1980, p. 96;
2007) as “[. . .] a summary of molar perceptions that employees
share about their work environments”, and particularly with regard
to safety policies, procedures, and practices. Similarly, Niskanen
(1994, p. 241) defines safety climate as “[. . .] a set of attributes that
can be perceived about particular work organizations and which
may be induced by the policies and practices that organizations
impose upon their workers.” Consistent with our theoretical and
operational definition of climate, we draw upon Neal et al. (2000)
definition of safety climate as a specific form of organizational cli-
mate, understood as individual perceptions of policies, procedures,
and practices relating to safety in the workplace.
Neal et al.’s (2000) conceptual definition of safety climate pro-
poses a higher order factor of organizational safety climate, which
incorporates four different dimensions. The first component refers
to members’ perceptions of the extent to which management is
concerned for employee well-being and safety (i.e., management
values). A second factor captures the perception of employees’
opportunities to discuss safety issues and concerns and whether
they are consulted about workplace health (i.e., safety commu-
nication). A third climate component refers to perception of the
adequacy of training in terms of comprehensiveness of safety issues
covered by the training and accessibility of safety programs (i.e.,
safety training). Finally, a fourth factor captures people’s perception
of the quality of safety management systems and their effective-
ness in preventive work incidents (i.e., safety systems). Measures
of safety climate (including Neal et al., 2000) tend to capture a
poor-to-positive perception of safety features (e.g., where lower
scores represent a poor safety climate and higher scores represent
a positive safety climate).
2.2. Safety culture
Despite the fact that safety culture commonly refers to organi-
zational schemas developed in order to deal with error and risk
management, the extant literature provides myriad approaches
to the topic (Choundry et al., 2007). According to Hale (2000),
safety culture refers to attitudes, beliefs, and perceptions shared
by natural groups that define norms and values, which in turn will
determine the way in which people will act and react to risk and
79
Organizational
Autocratic
Safety Culture Bureaucratic
RQ1 Clan-Patronage
Dimensions Technocratic
Cooperative
RQ2
RQ3
Safety
Compliance
H1
risk management. Cooper (2000) defines culture as the product of
multiple goal oriented interactions among people and situations. In
particular, safety culture is the observable extent to which all orga-
nizational members put their effort in improving safety on a daily
basis. More recently, Richter and Koch (2004) define safety culture
as shared and learned meanings. They argue that culture refers to
job and safety related experiences and the way they are interpreted
by people. In turn, this guides employee actions with regard to job
risk, accidents, and their prevention. Furthermore, Fang et al. (2006)
suggest that safety culture provides the organizations with a set of
prevailing indicators, beliefs, and values with regard to job safety.
As can be seen, many of these definitions contain aspects of both
organizational safety climate and safety culture.
Therefore, in an attempt to clearly distinguish safety culture
from safety climate, we relied on Schein’s (1999) conceptualization
of organizational culture to develop our definition of safety cul-
ture. Schein’s model incorporates three embedded levels of culture
expression. The most visible and external layer refers to organi-
zational artefacts (e.g., language, furniture, dress codes). The next
underlying level includes the organizational norms and values that
contribute to shape how artefacts are modelled and manifested. The
third and deepest cultural level refers to the basic assumptions that
members hold about their organizational reality and its function-
ing. These deep assumptions are so ingrained in the members’ mind
to the point that they may be less aware of their assumptions. Not
surprisingly, this constitutes the most stable part of organizational
culture and the one most resistant to change.
In alignment with Schein’s approach, we define safety culture
as a shared set of safety-related attitudes, behaviors, values, and
ingrained assumptions that orient organizational action pertaining to
safety. Although Schein (1985) suggests that culture can only be
assessed by a qualitative and context-specific examination of an
organization, we advocate that the values, norms and deep beliefs
proposed by his model are latent variables that psychology com-
monly measures with quantitative tools. Therefore, we maintain
that their measurement can be developed at the organizational
level as well (i.e., attitudes, behaviors, values, beliefs that members
of an organization have in common), and draw upon the cultural
stratification proposed by Schein in order to quantify the mea-
surement of each different cultural layer. Despite Guldenmund’s
(2007) skepticism regarding the ability to measure safety culture
using quantitative surveys, others (e.g., Mearns et al., 2013) demon-
strated the development and validity of a measure of safety culture
within the airline industry. Glendon and Stanton (2000, p. 3) further
provide guidance for the development of such culture measures,
noting“[. . .] key dimensions of organizational culture that have
been identified include depth, breadth and progression (Eldridge
and Crombie, 1974). Depth refers to the way in which culture
is reflected in the organization’s policies, programs, procedures,
80 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89
practices, values, strategies, behaviours and other features. Cul-
tural breath is represented in the lateral coordination of different
organizational components.” Consistently, Payne’s (2000) multi-
dimensional model of organizational culture provides conceptual
definitions of intensity and strength in line with the above contri-
butions, and also suggest how to operationalize both the layered
(i.e., intensity) and shared (i.e., strength of embracement among
members) features of organizational culture.
Recently, Petitta et al. (2014) and Petitta et al. (in press) pro-
posed and found empirical support for amodel of organizational
safety culturethat assesses the extent to which five cultural dimen-
sions (e.g., autocratic, bureaucratic, clan-patronage, technocratic,
and cooperative; Enriquez, 1970) are deeply rooted (i.e., intensity)
and the extent to which they are shared among organizational
members (i.e., strength). The autocratic safety culture dimension
assesses the extent to which one’s own direct superior/leader (i.e.,
authority) is the source of safety instructions and directions for
employees. Dialogue mainly consists of the delivery of safety direc-
tives and feedback involves corrections which highlight errors to
avoid. Within the bureaucratic safety culture dimension, the funda-
mental value is each members’ adherence to organizational safety
norms and regulations set by top level bureaucratic officials (i.e.,
experts who develop norms also in line with local government
requirements). Within the clan-patronage safety culture dimension,
powerful people within a group of privileged members provide
benefits that individuals seek to obtain and that are available to
reciprocate, thus contributing to create a context that works with
different rules depending on the in-group or out-group member-
ship. This “two-faced” context provides members with different
safety rules and directions dependent upon their current inter-
action with members of their inner circle versus more external
people. The technocratic safety culture dimension reflects contexts
which are results-oriented and focused on competition. However,
prioritizing achievement orientation may result in safety violations
if shortcuts to excellence include hiding errors, or skipping safety
steps, etc. Finally, the cooperative safety culture assesses the extent
to which value is given to the achievement of consensus and the
participation of all members, and an emphasis is placed on the
contributions of all individuals to collective safety outcomes.
It isimportant to note that these dimensions are not mutually
exclusive (Marradi, 1990). When assessing the culture of an organi-
zation, each of the dimensions included in the model is necessary to
fully describe/explain theorganizational safety culture. While dif-
ferent dimensions may coexist, one may prevail over the others
(Valsiner and Connolly, 2002).
As such, organizational safety culture provides a more qualita-
tive distinction about the types of ingrained safety-related beliefs,
attitudes, and values within the organization, whereas safety cli-
mate tends to capture perceptions of safety features along a
continuum ranging from poor to positive (i.e., where lower scores
represent a poor safety climate and higher scores represent a posi-
tive safety climate). Because this study represents the first study to
explicitly measure first-order safety culture dimensions (i.e., Petitta
et al., in press), along with higher-order safety climate (i.e., Neal
et al., 2000), we pose the following research question:
Research Question 1: How are the five organizational safety cul-
ture dimensions related to organizational safety climate?
3. Safety climate and safety culture as predictors of safety
compliance
Research on the role of organizational and psychosocial fac-
tors in influencing individual safety outcomes demonstrated that
safety climate has great impact on workers’ behavior. Meta-analytic
results (Clarke, 2006) generally suggest that a positive safety cli-
mate is associated with higher levels of safety participation and,
to a lesser extent, with more safety compliance. More recently,
Cavazza and Serpe (2009) found that the perception of a positive
safety climate has a direct and negative effect on unsafe behaviors,
thus improving workers compliance with safety norms. Similarly,
Kapp (2012) showed that under positive safety climate conditions,
employee safety compliance behavior improves, thus providing
further support to the value of strong safety climates for reducing
the likelihood of risk and danger behaviors.
Moving towards the role of safety culture as a predictor of safety
compliance, a recent study by Martínez-Córcoles et al. (2011) in
a nuclear plant suggested that safety culture values have a direct
impact on general safety behaviors. The study also provides support
to the direct role of safety climate in predicting safety behaviors.
Although Martínez-Córcoles et al.’s (2011) study simultaneously
investigates safety climate and safety culture as antecedents of
safety behaviors, their safety culture approach is a profiling and
unidimensional one (i.e., only distinguishing strong vs. weak cul-
ture). Furthermore, it focuses on safety behaviors at large rather
than on specific safety compliance.
In a study from a wide variety of organizations in at-risk industry
sectors, Petitta et al. (in press) found that specific cultural dimen-
sions(particularly bureaucratic and technocratic) are differentially
related to the enactment of safety-related moral disengagement
and subsequent accident underreporting. Although they did not
examine safety compliance, their study provides initial evidence
on how different safety culture types impact employee safety out-
comes, i.e., accident underreporting. Moreover, with regards to
Italian organizational contexts, safety climate has been found to be
positively associated with safety performance (Barbaranelli et al.,
2015; Brondino et al., 2012).
To date, we are not aware of any study examining the simul-
taneous direct effect of safety climate and different culture
dimensionson safety compliance behaviors. However, one recent
study conducted within the health care field indicates that it may
be important to consider their unique contributions. Specifically,
in their study of 30 different Veteran Affairs hospitals across the
United States, Hartmann et al. (2009) found evidence for four dif-
ferent culture types (entrepreneurial; hierarchical, group-focused,
and rational). Additionally, they found that the safety climate
exhibited within these hospitals was related to these different
culture types. Despite this initial intriguing evidence, researchers
acknowledge that little is still known regarding the relation-
ship between organizational culture, climate, and safety behaviors
(Ross, 2011). Hence, rather than develop specific hypotheses, we
pose the following research question:
Research Question #2: Will each of the five different cultural
dimensions (i.e., autocratic, bureaucratic, clan-patronage, techno-
cratic, and cooperative) conjointly measured with safety climate
differentially explain variance in safety compliance?
4. Safety climate and safety culture as moderators of the
relationship between supervisor enforcement and safety
compliance
Initial evidence suggests that a positive safety climate may
enhance the relationship between safety leadership and employee
safety behaviors (Kapp, 2012; Probst, 2015). Hofmann et al. (2003)
found that a positive safety climate enhanced the positive rela-
tionship between transformational safety leadership and safety
participation. Our study extends this scant literature by considering
the relationship between a transactional form of safety leadership
(i.e., supervisor safety enforcement) and safety compliance. Addi-
tionally, our study contributes to the extant literature, since we are
unaware of any study examining safety culture as a moderator of
 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89
the relationship between supervisor enforcement and safety com-
pliance. The only study to consider both culture and climate was
the Martínez-Córcoles et al. (2011) study. Unfortunately, the test of
their hypotheses was limited due to reliance on data drawn from
employees in a single organizational setting. Thus, they could not
explicitly test the effects of different cultures and/or safety climates
at the appropriate level of analysis.
Given that no study has previously examined the simultane-
ous moderating effects of safety climate and different culture
dimensionson the relationship between supervisor enforcement
and safety compliance, we pose the following research question:
Research Question #3: What are the conjoint influences of safety
climate and the five different cultural dimensionson the relation-
ship between supervisor safety enforcement and employee safety
compliance?
5. Method
5.1. Participants
Data were collected from 1342 employees in 32 Italian organiza-
tions (67.5% average response rate). Level 1 (individuals) and Level
2 (organizations) were both sampled by convenience. Together,
these organizations represent a wide range of industry sectors
where safety compliance is a paramount concern, including1 man-
ufacturing (4 organizations, N = 294), construction (7 organizations,
N = 245), transportation (4 organizations, N = 59), military (3 orga-
nizations, N = 185), energy (4 organizations, N = 279), health care (2
organizations, N = 142), and distribution/service (8 organizations,
N = 138). The organization’s size ranged from small to medium and
large. The mean organizational sample size was about 42 employ-
ees and ranged from 5 to 130. Overall, participants were 83.8%
males, the average age was 41.36 years (SD = 10.58), and the average
tenure in the position was 12.13 years (SD = 9.65). Fifty-five per-
cent were blue-collars, 30.9% were employees whereas 8.9% held
the role of supervisor.
5.2. Procedure
The research team approached administrators within each orga-
nization to request their organization’s participation in the study.
Upon reaching agreement on participation, the research team
provided information sessions at each organizational location to
describe the project, encourage participation, and address concerns
from potential participants. Participants were informed that the
study aimed to assess how workplace safety may be affected by the
organizational environment and policies. Participation was volun-
tary and anonymous. The research team distributed questionnaires
which the majority of participants completed that same day. In
some instances, employees were allowed up to two weeks to com-
plete the survey at home and return it in a sealed envelope to the
research team.
5.3. Study measures
The following scales were translated into Italian from the
English version using the standard translation-back-translation
procedure recommended by Brislin (1980): Safety Compliance,
1
In-depth sample demographics by organization are available upon request to
the first author. This sample also partially overlaps with Petitta et al. (in press);
however, the only variable these two studies share is the measure of safety culture.
Whereas the current study examines supervisor enforcement, safety climate, and
safety compliance, the Petitta et al. (2014) study examined moral disengagement
and accident underreporting.
81
Supervisor Enforcement, Safety Climate. The Organizational Cul-
ture Safety scale was originally developed in Italian.
5.3.1. Safety compliance
Safety Compliance was assessed using the Italian version
(Barbaranelli et al., 2015) of the Neal et al. (2000) 4-item scale
designed to assess the degree to which employees comply with
safety rules and policies (e.g., “I use the correct safety procedures
for carrying out my job”). Items were presented in Likert-type for-
mat with a scale ranging from strongly disagree (1) to strongly agree
(7).
5.3.2. Supervisor enforcement
Supervisor enforcement of safety policies was assessed with the
extrinsic safety motivation scale developed by Probst and Brubaker
(2001). This 4-item scale assesses the degree to which employees’
supervisors enforce and provide rewards based on compliance with
safety rules and policies (e.g., “My supervisor strictly enforces safety
rules and regulations”). Items were presented in Likert-type format
with a scale ranging from strongly disagree (1) to strongly agree (7).
5.3.3. Safety climate
Safety climate was measured using the Italian version
(Barbaranelli et al., 2015) of the Neal et al. (2000) 16-item scale
designed to assess four dimensions of safety climate including man-
agement values (e.g., “Management places a strong emphasis on
workplace health and safety”), safety communication (e.g., “There is
open communication about safety issues within this workplace”),
safety training (e.g., “Safety issues are given a high priority in
training programs”), and safety systems (e.g., “There are system-
atic procedures in place for preventing breakdowns in workplace
safety”). Items were presented in Likert-type format with a scale
ranging from strongly disagree (1) to strongly agree (7). Individual-
level responses can be aggregated to create an organizational-level
measure of climate by taking the mean of the individual employees’
climate scores within each organization (Neal and Griffin, 2006).
Higher scores are interpreted to reflect a more positive organiza-
tional safety climate.
5.3.4. Safety culture
To measure safety culture, we utilized the 20-item Job Safety
sub-scale of the Intensity & Strength Organizational Culture
Questionnaire (JS-I&SOCQ; Petitta et al., 2014; Petitta et al., in
press). This sub-scale measures the five cultural dimensions (auto-
cratic, bureaucratic, clan-patronage, technocratic, and cooperative)
within a safety context. Respondents are first provided the follow-
ing prompts describing each of the cultural dimensions with safety
as the frame of reference: “With respect to enacting safety behav-
iors that are indicated by the boss. . .” (autocratic); “With respect
to strictly adhering to the rules and safety procedures proposed
by the organization. . .” (bureaucratic); “With respect to following
one set of safety rules when you are with outsiders but following
different practices within the group. . .” (clan-patronage); “With
respect to following the safety practices only if this doesn’t impede
the achievement of the result and the progress of the work. . .”
(technocratic); and “With respect to proactively involving all mem-
bers of the organization in the diffusion and adoption of safety
practices. . .” (cooperative). After viewing the prompts, respon-
dents indicate how many people (ranging from 1 = almost no one
to 4 = almost everyone) from their organization: 1) manifest a posi-
tive attitude towards the cultural dimension, 2) engage in behavior
that is in line with the cultural dimension, 3) attribute importance
to the cultural dimension, and 4) consider the cultural dimensionso
fundamental as to consider it deeply ingrained in the organizational
82 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89
texture. Thus, as can be seen, the JS-I&SOCQ uses a referent-shift2
approach (Wallace et al., 2016) to the measurement of safety cul-
ture. As such, the JS-I&SOCQ asks respondents to indicate their
perception of how others in their context appraise and display some
work environment characteristics. Therefore, in the referent-shift
composite method the extent to which members’ perceptions are
shared at the organizational level is inherently provided by the focus
on others in the same context (i.e., by asking respondents to indicate
from 1 = “almost no one” to 4 = “almost everyone” how widespread
the perceptions are). This approach obviates the necessity of exam-
ining within-group agreement assessments (i.e., rwg ) as an indicator
of the inter-rater reliability of individually-held perceptions, as in
the direct consensus approach (Wallace et al., 2016).
The progressively deeper intensity of safety culture layers could
vary from low (attitudes) to high (deeply rooted belief), whereas
the strength of the safety culture is reflected in estimates of how
many people express that cultural feature/layer (e.g., almost no one
to almost everyone). The overall score of each culture dimension
(e.g., autocratic) is obtained by averaging the answers provided to
progressively deeper (intensity) culture layers (e.g., autocratic atti-
tude, autocratic behaviour, autocratic value, and autocratic deeply
rooted belief) within that specific culture type. Therefore, higher
mean scores of autocratic, bureaucratic, clan-patronage, techno-
cratic, and cooperative dimensions are interpreted to reflect a
greater manifestation (i.e., in terms of both intensity and strength)
of that cultural typology within the employees’ organization.
Previous research on the JS-I&SOCQ (Petitta et al., 2014, in press)
reported excellent psychometric properties with respect to the
scale’s reliability (dimension alphas ranging from 0.90 to 0.96).
Petitta et al. (in press) provide additional evidence of construct
validity and discriminant validity via confirmatory factor analy-
sis. Below we also provide an independent test of the discriminant
validity.
6. Results
6.1. Measurement model and reliability
In order to address the relationship between safety culture and
safety climate measures, we first examined a six-factor confirma-
tory factor analysis (CFA) model in which each item referring to the
culture dimensions (i.e., autocratic, bureaucratic, clan-patronage,
technocratic, and cooperative) respectively loaded onto five unique
latent factors, and each item referring to the four climate sub-scales
(i.e., management values, safety communication, safety training,
safety systems) loaded onto one unique second-order factor. Given
the multilevel or hierarchical structure of our data wherein indi-
viduals are nested within organizations, we used the MPLUS
“TYPE = COMPLEX” procedure of MPLUS (Muthén and Muthén,
1998–2012). This MPLUS command produces correct parameters
estimates, standard errors, and test statistics in presence of inter-
dependency. The model was tested on the covariance matrix, using
the Robust Maximum Likelihood (RML) estimation method. Sub-
jects with more than 50% of missing values were not considered
for these analyses. Furthermore, residuals pertaining to the same
2
It is important to note that there are two different approaches that scholars
have taken to the measurement of higher-level constructs, such as organizational
culture: a direct consensus approach and a referent shift approach. “Direct con-
sensus typically aggregates survey items that begin with an individual perception
(e.g., “I believe . . .”) whereas referent-shift consensus typically aggregates survey
items that reflect an individual’s perception of some higher-level structure (e.g.,
“My team believes . . .”).” (Wallace et al., 2016; p. 2). Therefore, “The shift in refer-
ent is accomplished by simply changing the referent in each survey item from “I”
(in direct consensus) to “we” or “the group” (in referent-shift consensus—or some
other indicator that refers to the higher level) (Wallace et al., 2016: p. 4).
item across the five factors reflecting different cultural dimensions
were allowed to correlate. We used the Akaike Information Cri-
terion (AIC; Akaike, 1987) and Expected Cross-Validation Index
(ECVI; Browne and Cudeck, 1993) for model comparison purpose
(Brown, 2015). Lower values of these indices indicate a better fit to
the observed data.
Next, in an attempt to demonstrate the conceptual distinctive-
ness of both safety culture dimensions and safety climate from the
constructs of supervisor enforcement and safety compliance, we
tested an eight-factor measurement model in which a) each item
that referred to supervisor enforcement, safety compliance, and the
five different safety cultural dimensions respectively loaded onto
seven unique latent factors, and b) the items that referred to the
four safety climate sub-scales loaded onto a second-order safety
climate factor. Results from the eight-factor CFA showed excellent
fit indices: ␹2 (830, N = 1303) = 2075.45, p < 0.001, RMSEA = 0.034
(.032; 0.036), CFI = 0.96, TLI = 0.96, with factor loadings all signifi-
cant and above 0.57. Correlations among climate and culture factors
ranged from −0.05 to 0.39, with a mean of 0.18 and a standard
deviation of 0.29. On the other hand, correlations among the five
culture factors ranged from −0.05 to 0.74, with a mean of 0.26 and
a standard deviation of 0.32. Finally, the Cronbach’s alpha for the
empirical scales ranged from 0.75 to 0.97 (see Table 1). Moreover,
all the values related to Composite Reliability (CR) and Average
Variance Extracted (AVE) met the criteria suggested by Raykov and
Marcoulides (2011) and were satisfactory. Together, these findings
demonstrated the appropriateness of the eight hypothesized latent
factors, the distinctiveness among the study constructs, and, in par-
ticular, between safety climate and the safety culture dimensions,
thus providing an initial answer to Research Question #1.
6.2. Descriptive statistics and correlations
The descriptive statistics and zero-order correlations among
our study’s variables3 were calculated at individual- and
organizational-level and suggest an interesting preliminary pattern
of relationships (see Table 1). Specifically, at the organizational-
level (and in line with previous research), we see that organizations
with a more positive safety climate have significantly higher levels
of supervisor enforcement (r = 0.79, p< 0.01) and higher levels
of employee safety compliance (r = 0.77, p< 0.01). Interestingly,
comporting with Hartmann et al. (2009), we also see that certain
cultural dimensions tend to have more positive safety climates and
outcomes than others. For example, autocratic, bureaucratic, and
cooperative cultures are positively correlated with safety climate
(r = 0.61, 0.64, and 0.67, respectively, all p< 0.01), thus providing
additional insight into RQ#1. These same cultural dimensions are
also correlated with more positive safety outcomes. For example,
autocratic, bureaucratic, and cooperative safety cultures were
related to more positive compliance and enforcement (r’s ranging
from 0.54 to 0.72). On the other hand, the technocratic culture
was negatively related to safety climate (r = −0.38, p< 0.05) and
supervisor enforcement (r = −0.58, p< 0.01), and exhibited the
same trend for compliance. Finally, the clan-patronage culture
was slightly related to safety compliance, and unrelated to safety
climate or supervisor enforcement. Together, these data provide
an initial test of Research Question #2 by suggesting that safety
compliance will be greater in more autocratic, bureaucratic, and
cooperative safety cultures (although the underlying mechanism
for the increased compliance may differ). On the other hand,
being employed in a clan-patronage culture does not appear to be
significantly related employee safety compliance levels, whereas
3
The full results from the comparison of alternative models is available upon
request to the first author.
 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89 83

Table 1
Means, Standard Deviations, and Zero-Order Correlations.

Variable M SD CR (AVE) 1 2 3 4 5 6 7 8

1. 1. Climate Total 4.67 1.26 0.98 (0.75) (0.97) 0.34** 0.33** −0.06* −0.20** 0.36** 0.60** 0.60**
2. 2. Autocratic Culture 2.90 0.74 0.88 (0.62) 0.61** (0.86) 0.70** 0.12** −0.03 0.55** 0.32** 0.32**
3. 3. Bureaucratic Culture 2.96 0.73 0.89 (0.63) 0.64** 0.84** (0.87) 0.10** −0.04 0.60** 0.32** 0.27**
4. 4. Clan-Patronage Culture 2.37 0.90 0.93 (0.75) −0.04 0.17 0.24 (0.92) 0.51** 0.06* −0.09** 0.01
5. 5. Technocratic Culture 2.11 0.89 0.94 (0.77) −0.38* −0.25 −0.20 0.70** (0.93) −0.04 −0.22** −0.06*
6. 6. Cooperative Culture 2.79 0.79 0.92 (0.72) 0.67** 0.77** 0.86** 0.37* −0.04 (0.89) 0.32** 0.35**
7. 7. Safety Compliance 5.31 1.23 0.93 (0.74) 0.86** 0.61** 0.66** 0.06 −0.23 0.72** (0.92) 0.46**
8. 8. Supervisor Enforcement 3.95 1.15 0.74 (0.39) 0.73** 0.69** 0.69** −0.23 −0.58** 0.54** 0.63** (0.75)

Note: * p < 0.05; **p < 0.01. Correlations above the diagonal are at the individual-level (listwise n = 1278), whereas those below the diagonal are at the organizational-level
(N = 32). Means and SD are reported at the individual-level; Cronbach’s alpha at the individual-level is reported along the diagonal in parentheses. CR = Composite Reliability;
AVE = Average Variance Extracted.

a technocratic culture appears to have negative effects on safety
compliance.
6.3. Hierarchical linear modeling
As indicated by Bliese (2000), we first assessed the intra-
class correlations (ICC [1]) and the reliability of the means (ICC
[2]) to evaluate the appropriateness of aggregating safety cli-
mate and safety culture dimensions scores to the organizational
level. The ICC (1) was 0.08 for safety climate, F (31,1310) = 19.22,
p < 01;.12for autocratic culture, F (31,1267) = 6.71, p < 001; 0.12
for bureaucratic culture, F (31,1270) = 6.54, p < 001; 0.06for clan-
patronage culture, F (31,1261) = 3.60, p < 001; 0.12for technocratic
culture, F (31,1263) = 6.44, p < 001;and 0.08 for cooperative culture,
F (31,1261) = 4.87, p < 001. Although some of the ICC(1) values are
lower than the recommended ≥ 0.12 (James, 1982), they were all
statistically significant, thus justifying aggregation (Bliese, 2000).
The ICC (2) was 0.80, 0.85, 0.85, 0.72, 0.84, 0.80 for safety-
climate, autocratic culture, bureaucratic culture, clan-patronage
culture, technocratic culture, and cooperative culture, respectively,
meeting the recommended guidelines of ≥0.60 (Bartko, 1976;
Shrout and Fleiss, 1979). Moreover, following Heck and Thomas’
(2015) recommendations, we computed the Design Effect index
(Deff; Muthen and Satorra, 1995) in order to further assess the
between-organization variability for safety climate and safety cul-
ture dimensions. As suggested by Muthen and Satorra (1995), a
Deff > 2 reflects a non ignorable interdependence of observations
within organizations. In our study, Deff values ranged from 3.46
(paternalistic) to 5.92 (autocratic, bureaucratic, technocratic, and
cooperative) for safety culture, whereas Deff for safety climate was
4.28, thus indicating that organizational variables (i.e., safety cul-
ture and safety climate) could be appropriately aggregated at level
2.
In order to formally test our hypotheses to determine
whether supervisor enforcement predicts safety compliance (i.e.,
Level 1–individual-level variables) and whether organizational
safety climate and the organizational culture dimensions (i.e.,
Level 2–organizational-level variables) moderate the relationship
between supervisor enforcement and safety compliance, we chose
hierarchical linear modeling to analyze the data (Raudenbush and
Bryk, 2002). All predictors were centered at the grand mean and
parameters were estimated using restricted maximum likelihood
estimation (REML; Enders and Tofighi, 1998; Hofmann and Gavin,
1998). In particular, we tested the simultaneous moderating effect
of organizational climate and culture. Safety climate was consid-
ered as a unidimensional second-order variable including the four
sub-factors of management values, safety communication, safety
training, and safety systems (Neal et al., 2000). Conversely, because
organizational culture consists of five unique dimensions, we pro-
ceeded as follows. A preliminary null model assessed whether
sufficient variance between organizations allowed us to continue
testing our hypotheses (i.e., ICC1 ; Bliese, 2000; LeBreton and Senter,
2008). We then ran a Model 1 that included individual-level super-
visor enforcement and safety compliance. Next, Model 2 added
organizational-level safety climate and safety culture dimensions
in order to explain safety enforcement’s intercepts variation. In
particular, because organizational culture is a five-factor construct,
Model 2 was run five separate times, once for each of the different
culture dimensions (autocratic, bureaucratic, clan-patronage, tech-
nocratic, and cooperative, respectively). As an illustrative example,
the level-1 and level-2 equations for the test of the autocratic
dimension Model 2 are shown below:
Level 1: Yij (Safety Compliance) = ˇ0j + ˇ1j (Supervisor
Enforcementij ) + rij
Level 2 (Autocratic Culture): ␤0j = ␥00 + ␥01 (Safety
Climatej ) +  02 (Autocratic Culturej ) + U0j
␤1j =  10 + U1j
where Yij is the observed value of the outcome variable for respon-
dent i nested within organization j, ␥00 is the level-2 intercept (i.e.,
the grand mean of compliance), ␥01 and␥02 are the level-2 slopes,
␥10 is the mean slope corresponding to the level-1 predictor, rij
is the within-organization variance in compliance, U0j is the
residual intercept variance, and U1j is the variance in slope asso-
ciated with the level-1 predictor.
Finally, Model 3 included the cross-level supervisor
enforcement × safety climate and supervisor enforce-
ment × organizational culture dimensions terms. Similar to
Model 2, Model 3 was run five separate times (once for each
culture dimension), each time including the supervisor enforce-
ment × safety climate interaction term and the supervisor
enforcement × organizational culture dimensions interaction
(i.e., autocratic, bureaucratic, clan-patronage, technocratic, and
cooperative) as cross-level interactions. Below is a sample equation
for Model 3:
Mixed Model(Autocratic Culture) : Y ij (Safety Complianceij)
=  00 +  01 (Safety Climatej ) +  02 (Autocratic Culturej )
+  10 (Supervisor Enforcementij ) +  11 (Safety Climatej
∗Supervisor Enforcementij ) +  12 (Autocratic Culturej
∗Supervisor Enforcementij ) + u0j + u1j (Supervisor
Enforcementij ) + rij
where Yij is the observed value of the outcome variable for respon-
dent i nested within organization j, ␥00 is the level-2 intercept,
␥01- ␥02 are the level-2 slopes, ␥10 is the mean slope correspond-
ing to the level-1 predictor, ␥11- ␥12 are the cross-level interaction
84 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89

Fig. 2. Cross-level effect of autocratic safety culture on the supervisor enforcement-

safety compliance relationship (cross-level interaction of safety climate is not
significant). Fig. 3. Cross-level effect of bureaucratic safety culture on the supervisor
enforcement-safety compliance relationship (cross-level interaction of safety cli-
mate is not significant).
terms, U0j is the residual intercept variance, and U1j is the residual
variance in slope associated with the level-1 predictor.
The test of the null model indicated that organizational differ-
ences accounted for 19.05% of the total variance in the proportion
of safety compliance, suggesting sufficient variance between orga-
nizations to continue testing our hypotheses (LeBreton and Senter,
2008).
As can be seen in Table 2, higher supervisor enforcement was
related to significantly higher safety compliance (␥10 = 0.37, t = 9.92,
p< 0.001). Even after adding the level-2 variables in Models 2 and 3,
the main effect at level-1 persisted. Thus, we found strong support
for Hypothesis 1.
Although we used the zero-order correlations to initially explore
Research Question #2, Table 2 provides a more rigorous test of
the level-2 main effects of organization safety climate combined
with the five different safety culture dimensions. In particu-
lar: (1) both the autocratic safety culture (␥02 = 0.42, t = 2.097,
p< 0.05) and safety climate (␥01 = 0.41, t = 6.621, p< 0.001) posi-
tively predicted safety compliance; (2) both bureaucratic culture Fig. 4. Cross-level effect of safety climate on the supervisor enforcement-safety
compliance relationship (cross-level interaction of clan-patronage safety culture is
(␥02 = 0.45, t = 3.121, p< 0.01) and the organization’s safety cli-
not significant).
mate (␥01 = 0.41, t = 9.306, p< 0.001) positively predicted safety
compliance; (3) the clan-patronage culture was not significant
(␥02 = −0.32, t = −1.673, p = 0.105), whereas organization’s safety p = 0.835), with the cross-level interaction explaining 12.37% of the
climate positively predicted safety compliance (␥01 = 0.46, t = 8.106, variance in the slopes of the enforcement-compliance relation-
p< 0.001); (4) the technocratic culture negatively predicted safety ship; (4) neither attenuated by the organization’s safety climate
compliance (␥02 = −0.42, t = −2.535, p< 0.05), whereas organi- (␥11 = −0.08, t = −1.883, p = 0.070), nor by technocratic culture
zation’s safety climate positively predicted safety compliance (␥12 = 0.08, t = −0.806, p = 0.427); (5) neither attenuated by the
(␥01 = 0.40, t = 6.131, p< 0.001); (5) the cooperative culture was not organization’s safety climate (␥11 = −0.06, t = −1.855, p = 0.074), nor
significant (␥02 = 0.24, t = 1.176, p = 0.249), whereas organization’s by cooperative culture (␥12 = −0.10, t = −1.072, p = 0.293).
safety climate positively predicted safety compliance (␥01 = 0.44,
t = 7.454, p< 0.001). 7. Discussion
Finally, we explored Research Question #3 by testing in Model
3 the extent to which the Level-1 relationship between supervisor In this study, we set out to: 1) examine the main effects of
enforcement and safety compliance was moderated by organi- climate and culture on employee safety compliance by utiliz-
zation safety climate combined with the five different safety ing a newly developed measure of organizational safety culture;
culture dimensions. As can be seen in Table 2, the supervisor and 2) evaluate the extent to which climate and culture might
enforcement-safety compliance relationship was: (1) attenuated conjointly influence the strength of the relationship between
by the organization’s autocratic culture (␥12 = −0.24, t = −2.365, supervisor enforcement of safety policies and employee safety
p< 0.05), but not by safety climate (␥11 = −0.03, t = −0.900, compliance behaviors. In doing so, we developed a model illustrat-
p = 0.375), with the cross-level interaction explaining 25.35% of ing the predicted relationships among these variables and tested
the variance in the slopes of the enforcement-compliance relation- our hypothesis and research questions within a multi-level model-
ship (see Fig. 2); (2) attenuated by the organization’s bureaucratic ing context using data drawn from 1342 employees nested within
culture (␥12 = −0.33, t = −2.738, p< 0.05), but not by safety climate 32 different organizations.
(␥11 = −0.01, t = −0.169, p = 0.867), with the cross-level interaction Consistent with prior research, the results indicated that: 1)
explaining 36.55% of the variance in the slopes of the enforcement- supervisor enforcement is significantly related to employee safety
compliance relationship (see Fig. 3); (3) slightly attenuated by the compliance; and 2) safety climate shows a significant main effect
organization’s safety climate (␥11 = −0.10, t = −2.619, p< 0.05; see on compliance. However, with regard to the latter point, our study
Fig. 4), but not by clan-patronage culture (␥12 = 0.03, t = −0.210, is the first to demonstrate that safety climate has a direct effect

on employee compliance under each of the five different safety
cultural dimensions. Adding to the existing literature, our study
also demonstrates that: 1) safety culture is a distinct construct
from safety climate; 2) knowledge of the organizational safety
culture can provide significant incremental information regarding
employee safety compliance, over and above supervisor enforce-
ment behaviors and the safety climate; and 3) certain cultural
Table 2
Multilevel Estimates for HLM Models.
Parameters
Constant
Level 1
Supervisor
Enforcement
Level 2
Safety Climate & [Autocratic Culture]
Safety Climate & [Bureaucratic
Culture]
Safety Climate & [Clan-Patronage
Culture]
Safety Climate & [Technocratic
Culture]
Safety Climate & [Cooperative
Culture]
Cross-Level Interaction
Safety Climate × Supervisor
Enforcement [Autocratic
Culture × Supervisor Enforcement]
Safety Climate × Supervisor
Enforcement [Bureaucratic
Culture × Supervisor Enforcement]
Safety Climate × Supervisor
Enforcement [Clan-Patron.
Culture × Supervisor Enforcement]
Safety Climate × Supervisor
Enforcement [Technocratic
Culture × Supervisor Enforcement]
Safety Climate × Supervisor
Enforcement [Cooperative
Culture × Supervisor Enforcement]
Variance Components
L-1 (r)
L-2 (u0)
Random Coefficiets
L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89
Null model Model 1
5.204 (0.103)*** 5.209
(0.102)***
– 0.365(0.037)***
– –
– –
– –
– –
– –
– –
– –
– –
– –
– –
1.186 0.973
0.286 0.296
– 0.025
dimensions (i.e., autocratic and bureaucratic) appear to suppress
the effect of safety climate and to attenuate the relationship
between supervisor enforcement and employee compliance, such
that high levels of compliance are seen regardless of supervisor
enforcement when these culture dimensions are strong within an
organization.
Model 2
5.188 (0.059)*** (climate + autocratic)
5.191 (0.057)***
(climate + bureaucratic)
5.205 (0.063)***
(climate + clan-patron.)
5.202 (0.061)***
(climate + technocratic)
5.193 (0.062) ***
(climate + cooperative)
0.339 (0.036)*** (climate + autocratic)
0.339 (0.036)***
(climate + bureaucratic)
0.355 (0.036)***
(climate + clan-patron.)
0.351 (0.036)***
(climate + technocratic)
0.343 (0.036) ***
(climate + cooperative)
0.413 (0.062)*** [0.415 (0.198) *]
0.412 (0.044)*** [0.445 (0.143) *]
0.458 (0.056) *** [−0.324 (0.193)]
0.396 (0.065) *** [−0.42 (0.165) *]
0.438 (0.058) *** [0.238 (0.202)]
–
–
–
–
–
0.0973 (climate + autocratic)
0.0975 (climate + bureaucratic)
0.0972 (climate + clan-patron.)
0.0971 (climate + technocratic)
0.974 (climate + cooperative)
0.086 (climate + autocratic)
0.078 (climate + bureaucratic)
0.103 (climate + clan-patron.)
0.09 (climate + technocratic)
0.101 (climate + cooperative)
0.023 (climate + autocratic)
0.023 (climate + bureaucratic)
0.023 (climate + clan-patron.)
0.022 (climate + technocratic)
0.023 (climate + cooperative)
85
Model 3
5.184 (0.059)*** (climate + autocratic)
5.191 (0.056)***
(climate + bureaucratic)
5.203 (0.062)***
(climate + clan-patron.)
5.198 (0.06)*** (climate + technocratic)
5.191 (0.061) ***
(climate + cooperative)
0.361 (0.036)*** (climate + autocratic)
0.361 (0.034)***
(climate + bureaucratic)
0.367 (0.035)***
(climate + clan-patron.)
0.363 (0.034)***
(climate + technocratic)
0.357 (0.036) ***
(climate + cooperative)
0.435 (0.075)*** [0.644 (0.231)**]
0.408 (0.068)*** [0.755 (0.214)***]
0.57 (0.066)*** [−0.413(0.253)]
0.482 (0.065) *** [−0.538 (0.173) **]
0.51(0.076) *** [0.304 (0.245)]
0.097 (climate + autocratic)
0.0987 (climate + bureaucratic)
0.0986 (climate + clan-patron.)
0.097 (climate + technocratic)
0.973 (climate + cooperative)
0.082 (climate + autocratic)
0.074 (climate + bureaucratic)
0.097 (climate + clan-patron.)
0.096 (climate + technocratic)
0.099 (climate + cooperative)
0.017 (climate + autocratic)
0.015 (climate + bureaucratic)
0.02 (climate + clan-patron.)
0.084 (climate + technocratic)
0.02 (climate + cooperative)
86 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89

Table 2 (Continued)

Null model Model 1 Model 2 Model 3

Model Fit Statistics

Deviance Statistic/No. 4103.652/2 3850.065/4 3823.198/4 (climate + autocratic)
Est. Parameters (df) 3822.952/4 (climate + bureaucratic)
3823.951/4 (climate + clan-patron.)
3820.899/4 (climate + technocratic)
3825.433/4 (climate + cooperative)
3824.329/4 (climate + autocratic)
3822.358/4 (climate + bureaucratic)
3826.234/4 (climate + clan-patron.)
3823.036/4 (climate + technocratic)
3828.457/4 (climate + cooperative)

Model Fit Tests

Diff Dev Stat (Diff df)
– 253.587(2)***
280.454(2)*** (climate + autocratic)
280.7(2)***(climate + bureaucratic)
279.701(2)*** (climate + clan-patron.)
282.753(2)***(climate + technocratic)
278.219(2)***(climate + cooperative)
279.323 (2)*** (climate + autocratic)
281.294 (2)***(climate + bureaucratic)
277.418 (2)*** (climate + clan-patron.)
280.616 (2)***(climate + technocratic)
275.195 (2))*** (climate + cooperative)

Note: Standard Errors are in parentheses; * p < 0.05, ** p < 0.01, *** p < 0.001; clan-patron.=clan- patronage.

Some additional intriguing results were the differential corre-
lations between different safety culture types and safety climate
at the organizational level. This suggests that perhaps organiza-
tions with particular safety cultures are more likely to develop
more (or less) positive safety climates. Safety climate is nearly uni-
versally measured as something ranging from poor to positive;
yet, safety culture provides a more qualitative distinction about
the ingrained safety-related beliefs, attitudes, and values within
the organization. Our current results suggest that certain types of
beliefs, attitudes, and values are more likely to be associated with
a positive safety climate (e.g., cooperative, bureaucratic, and auto-
cratic); others with a negative safety climate (technocratic); and
yet others (e.g., clan-patronage) are not associated with a positive
or negative safety climate. In addition, while a more positive safety
climate is related to better safety compliance, this is not uniformly
the case for all safety culture dimensions. Specifically, autocratic,
bureaucratic, and cooperative dimensions are positively related to
compliance. However, clan-patronage is not related to compliance,
whereas a stronger technocratic culture is correlated with less com-
pliance. We would assert that this differential patterns of results
helps to provide some discriminant validity evidence between the
two constructs of safety climate and culture, and should prompt
further research regarding the impact of safety climate within the
context of these different cultural dimensions.
Together, these results have theoretical implications for
researchers investigating contextual variables related to workplace
safety, as well as practical implications for practitioners striving to
improve safety behavior and outcomes among their workforce.
7.1. Theoretical contributions and implications
Although much attention has been paid to the economic and
social costs arising from safety breaches at work, no study has
previously examined how organizational safety climate and safety
culture simultaneously influence the extent to which employees
will comply with safety guidelines, nor how these variables might
influence the established individual-level relationship between
supervisor enforcement behaviors and employee compliance. Fur-
thermore, despite the fact that scientists and practitioners call for
clear theoretical and measurement distinctions between climate
and culture constructs, the two terms are often used interchange-
ably (Cox and Flin, 1998; Mearns and Flin, 1999; Yule, 2003).
Moreover, many purported studies and measures of culture appear
to be tapping into safety climate (Yule, 2003).
Therefore, we contributed to the current literature by: (a) pro-
viding a differential theoretical and operational background for
safety climate and culture; and, (b) using a multilevel modelling
approach to appropriately investigate the extent to which the
individual-level relationship between supervisor enforcement and
safety compliance may be moderated by organizational-level safety
climate and safety culture. Below, we discuss in greater detail the
theoretical and empirical significance of this study.
First, the current research contributes to extant literature
by conceptually and empirically disentangling the constructs of
organizational climate and culture that have been fraught with con-
ceptual ambiguity and overlap. As discussed in the Introduction,
their equivocal conceptualization and measurement has resulted
in a flourish of empirical research without the necessary accom-
panying theoretical development of the constructs. However, the
results of the multilevel CFA provided clear empirical support for
the conceptual difference between safety culture dimensions and
safety climate.
Second, the findings from our multilevel model examining the
conjoint main effects of culture and climate on employee safety
compliance, as well as the conjoint moderating effects of culture
and climate on the supervisor enforcement-safety compliance rela-
tionship, support the relevance of an expanded holistic approach
for studying safety at work. Although scholars have extensively
discussed the concepts of culture and climate (Ashkanasy et al.,
2010; Cooper et al., 2001), the ongoing research has tended to alter-
natively focus on either safety culture or safety climate. Hence,
our culture and climate framework serves as a complementary
contribution to ongoing research and represents the first attempt
to consider (from conceptual and operational perspectives) their
unique contributions to the understanding of employee safety. For
example, the study of the unique and conjoint effects of climate
and culture in the current study allowed us to identify the specific
organizational safety culture dimensions under which the effect
of safety climate is suppressed, and vice-versa. This information
allows for a more nuanced understanding of the conditions under
which contextual conditions of supervisor enforcement might
boost employee safety compliance.
Specifically,we found that both organizational-level autocratic
culture and climate positively predicted individual-level safety
compliance behaviors. However, the cross-level interaction of cli-
mate was suppressed by the autocratic culture, which attenuated
the supervisor enforcement-safety compliance relationship. Hence,
within an autocratic safety culture context, employees’ compliance
to safety rules is enhanced by supervisor enforcement, by a posi-
tive safety climate, and by a strong pattern of widespread attitudes,
behaviors, values, and beliefs suggesting to enact safety behaviors
that are indicated by the boss (i.e., autocratic culture). Nonetheless,
within a strong autocractic culture, the injunctive role of the super-
visor in boosting employee adherence to safety requirements does
not appear to be influenced by safety climate. We note that some lit-
erature findings address the problematic nexus between autocratic
leadership styles and workplace safety (Bohan and Laing, 2012; Van
De Vliert and Einarsen, 2008); yet, they refer to leadership models
rather than organizational culture and also apply to individual-
level managerial behaviours (see also GLOBE project; House et al.,
 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89
2004) rather than to organizational-level culture patterns, as is
the case of our study. Furthermore, a study from Su et al. (2012)
defined organizational-level autocratic culture as an obedience and
power based pattern of authoritarianism resulting in deterioration
of employees’ work environment in healthcare settings. Hence, the
focus of autocratic models is mainly on individual-level leadership
rather than on organizational-level culture patterns, and the differ-
ential conceptualization of our autocratic culture model makes our
findings an innovative contribution to this stream of safety enquiry.
Similarly, when concurrently examining bureaucratic safety
culture and safety climate, both positively predicted safety com-
pliance behaviors. However, even in this case, the cross-level
interaction of climate was suppressed by the bureaucratic culture,
which attenuated the supervisor enforcement-safety compliance
relationship. Hence, within a bureaucratic safety culture context,
employees’ compliance to safety rules is enhanced by supervisor
enforcement, by a positive safety climate, and by a widespread
pattern suggesting to strictly adhere to the rules and security proce-
dures proposed by the organization (i.e., bureaucratic culture). Still,
the injunctive role of the supervisor in accomplishing employees’
safety compliance does not appear to be influenced by safety cli-
mate, which was bounded by bureaucratic safety culture patterns.
In addition to proposing different conceptualizations of
bureaucratic culture, previous safety studies on the role of
organizational-level bureaucratic culture have also provided incon-
sistent findings. Chen et al. (2012) found that safety behaviors of
hospital staff are partly influenced by the prevailing bureaucratic
cultural norms in their organizations and work. On the other hand,
Speroff et al. (2012) found that a bureaucratic organizational cul-
ture, understood as a hierarchical model, is less favourable to safety
and quality improvement. Again, our differential conceptualiza-
tion of the bureaucratic culture and its conjoint study with safety
climate make our findings an innovative contribution.
Interestingly, when simultaneously investigating clan-
patronage safety culture and safety climate, only safety climate and
supervisor enforcement positively predicted safety compliance
behaviors. Furthermore, only a positive safety climate attenu-
ated the supervisor enforcement-safety compliance relationship.
Hence, both the direct and moderating effects of a clan-patronage
culture were suppressed by safety climate. The potential gain
from including the clan-patronage safety culture is the theoretical
and practical enrichment of the models made available in order
to comprehensively describe the culture patterns of an organiza-
tional context. Keyzer (2002) mentions paternalistic motives with
regard to product safety, yet from an economic perspective. Hence,
theclan-patronage dimension allows one toidentify a cultural
dynamic that has been neglected in earlier theoretical discussions
of culture, i.e., a clear distinction between who is inside and who
is outside of a “privileged” group, and the perception that informal
and unseen factors influence organizational behaviour.
Even more interestingly, when simultaneously investigating
technocratic safety culture and safety climate, the former nega-
tively predicted safety compliance behaviors. Specifically, within
a technocratic culture, employee compliance with safety rules
was enhanced by supervisor enforcement and by a positive safety
climate, whereas it was discouraged by a technocratic culture.
However, the supervisor enforcement-employee compliance rela-
tionship does not appear to be moderated by safety climate nor by
technocratic safety culture.
Literature contributions on technocratic models span from the
role of technocracy in the risk governance of national health care
(Kim, 2012), to the linear and technocratic knowledge-translation
approach to tackle complex public health problems (Jansen et al.,
2012). Despite the abundant reference to technocratic models, the
literature has not converged on a unique definition. Our culture and
climate framework contributes to the safety literature by disclosing
87
the isomorphic nature of the technocratic culture dimension with
the concept of production pressure (Probst et al., 2013) and their
similar detrimental effects on safety behaviors.
Finally, some surprising results emerged when simultaneously
investigating cooperative safety culture and safety climate. Specif-
ically, within a cooperative culture, employee compliance with
safety rules was enhanced by supervisor enforcement and by a
positive safety climate, whereas the effect of cooperative safety
culture is suppressed by safety climate. Furthermore, the relation-
ship between supervisor enforcement and employee adherence
to safety requirements was neither moderated by safety climate
nor by cooperative safety culture. Although differential coopera-
tive culture models have been proposed in the literature in order
to describe organizational culture patterns (Allen, 2006; Lloyd and
Carson, 2005; Phuong-Mai et al., 2005), there has been little empir-
ical inquiry on its relationship with safety prior to this study.
Overall, the present study demonstrated that supervisor
enforcement is particularly important to ensuring employee safety
compliance when an organization displays the following combi-
nation of contextual conditions: a clan-patronage or a cooperative
safety culture and a poor safety climate. Furthermore, within the
context of a technocratic safety culture and a poor safety cli-
mate, employee compliance requires supervisor enforcement in
order to contrast unwanted production pressure effects of a tech-
nocratic culture. Conversely, autocratic and bureaucratic safety
culture environments appear to render less important the roles of
safety climate and supervisor enforcement. Within these contexts,
safety compliance was high even in the face of poor climate and/or
enforcement.
Lastly, this study has implications for leadership at both the
supervisory and upper management levels by identifying the
effects of supervisor behaviors and safety climate within the
context of different safety cultures. This builds upon the extant lit-
erature, which previously suggested the differential role of micro
(supervisor) level factors and more macro level factors such as
upper-management actions as determinants of safety outcomes
(e.g., Simard and Marchand, 1997; Zohar, 2000, 2002; Zohar and
Luria, 2003, 2005). Taken as a whole, our findings encourage a mul-
tilevel and holistic approach that takes into account both culture
and climate in order to comprehensively understand the multi-
ple individual and contextual factors that shape safety compliance
behaviors.
7.2. Implications for practice
This study suggests initial practical implications for developing
effective means of improving employee compliance and thereby,
reducing unwanted organizational safety outcomes such as work-
place injuries. By better understanding the conjoint influence of
safety culture dimensions and safety climate on employees’ com-
pliance behaviors, and their link with the supervisor enforcement,
organizations may be able to more selectively focus their safety
improvement efforts (e.g., choosing between programs of safety
culture change, safety climate initiatives related to better train-
ing, clearer communication, etc., or focusing on supervisor-level
enforcement behaviors). Below we suggest a few possible practical
implications; however, it is important to note that before making
these “wholesale” recommendations, it is necessary to replicate
and extend the current findings.
For example, in organizations with poor safety climates, and a
clan-patronage, cooperative, or technocratic safety culture, super-
visor enforcement appears to be an important tool in engaging
employees’ conformity to safety. However, in organizations with
strong autocratic or bureaucratic cultures, employees’ compliance
is high regardless of the individual-level of supervisor enforcement.
From a practical perspective, this suggests that within the autocrac-
88 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89
tic or bureaucratic cultures, organization-wide efforts to improve
the overall safety climate and/or long-term programs for shaping
safety culture patterns may be more effective than targeting the
individual-level enforcement behaviors of supervisors. In particu-
lar, if the organization’s safety climate is a positive one, or the safety
culture autocratic or bureaucratic dimensions are strong, then the
extent to which employees fulfil safety requirements appears to be
largely independent of the enforcement behaviors of their individ-
ual supervisors. Conversely, individual-level interventions aimed
at empowering supervisors’ safety-related motivating leadership
style may be recommended in organizations with organizational-
level technocratic or cooperative safety cultures (in which safety
climate also failed to display a moderating role).
Managers should be made aware that employee variance in
safety-related behaviors depends not only on the organizational
safety climate, but also on the concurrent culture dimension(s) that
may beprevalent in their context. Hence, the use of the same safety
procedure may have different effects on employee behavior under
different organizational culture patterns. Our holistic and conjoint
safety culture and climate framework suggest that organizations
might benefit from including both a culture dimensions mapping
and safety climate evaluation in their safety surveys.
In particular, managers should realize that interventions that
aim to improve safety climate can focus on specific safety-related
contextual dimensions and how members portray them. Con-
versely, safety-related programs aimed at restructuring culture
patterns generally suggest to focus on changing programs that crys-
tallize the new organization’s values and behaviors (DeJoy, 2005).
Based on our findings, we suggest that managers should follow up
interventions’ effectiveness and make sure that the safety culture
dimension of the organization has been impacted and changed in
any significant component, namely attitudes, behaviors, values, and
beliefs.
7.3. Strengths, limitations and future perspectives
Although this study makes important contributions to the lit-
erature, it also suffers from several limitations that should be
addressed in future research efforts on this topic. First, although
the current set of data was drawn from numerous organizational
samples representing a wide variety of industry sectors (as well as
public/private and profit/non-profit), they were nonetheless con-
venience samples. Therefore, it is unclear if self-selection biases
in the kinds of organizations that agree to participate may have
affected our findings.
Second, our study relied on cross-sectional self-report data; as
such, it is possible that common method bias may threaten the
validity of the findings at the individual-level of analysis. However,
this would not have affected our primary hypotheses and questions
regarding the level-2 effects of climate and culture. Furthermore,
the simultaneous existence of high and low correlations among
the study variables suggests that common method bias unlikely
occurred in our study (Spector, 2006). Nevertheless, multi-source
and longitudinal data would better allow researchers to more rig-
orously assess the extent to which culture can be changed, the
expected time horizon for making such fundamental changes to
the fabric of an organization, and the extent to which these changes
result in employee safety improvements.
Finally, some scholars (e.g., Glendon and Stanton, 2000; Schein,
1985) argue that focusing solely on management behavior is
insufficient for ascertaining an organizational culture. Rather, it
is important to assess norms, values, and beliefs. A strength of
using the JS-I&SOCQ in the current paper is that its items directly
assess the attitudes, behaviour, values, and ingrained beliefs per-
taining to several different cultural dimensions (e.g., autocratic,
bureaucratic, technocratic) that are shared among organizational
members rather than solely focusing on managerial behaviors.
The JS-I&SOCQ also acknowledges that each organization may
have unique profiles with respect to these different safety culture
dimensions (e.g., high bureaucratic with low technocratic vs. high
autocratic and high bureaucratic). Future research examining the
relationships between these differing profiles and safety-related
outcomes would be the logical next step and better reflects the
complex reality of safety within the context of organizational
culture. However, although the five cultural dimensions of the JS-
I&SOCQ are measured each using 4 items (i.e., attitude, behaviour,
value, and ingrained belief) to capture the intensity of that dimen-
sion, each dimension only has one exemplar associated with it.
Thus, future improvements to the scale might include additional
exemplars of each safety culture dimension.
Despite these limitations, the current research is the first to pro-
pose an expanded holistic approach for studying safety at work
by simultaneously examining the effects of different safety culture
dimensions and safety climate on employee safety compliance, as
well as their moderating role on the supervisor enforcement-safety
compliance relationship. Hopefully, the results of our investigation
will assist future researchers and practitioners in better under-
standing the complex relationship between safety culture, safety
climate and the contextual circumstances under which supervisor
enforcement boosts employee safety compliance.
References
Akaike, H., 1987. Factor analysis and AIC. Psychometrika 52, 317–322.
Allen, L.Q., 2006. Investigating culture through cooperative learning. Foreign Lang.
Ann. 39 (1), 11–21.
Ashkanasy, N., Wilderom, C.P.M., Peterson, M.F., 2010. The Handbook of
Organizational Culture and Climate. Sage, Thouthand Oaks, CA.
Barbaranelli, C., Petitta, L., Probst, T.M., 2015. Does safety climate predict safety
performance in Italy and the USA? Cross-cultural validation of a theoretical
model of safety climate. Accid. Anal. Prev. 77, 35–44.
Bartko, J.J., 1976. On various intraclass correlation reliability coefficients. Psychol.
Bull. 83 (5), 762–765.
Bliese, P.D., 2000. Within-group agreement, non-independence, and reliability:
implications for data aggregation and analysis. In: Klein, K.J., Kozlowski, S.W.J.
(Eds.), Multilevel Theory, Research, and Methods in Organizations:
Foundations, Extensions, and New Directions. Jossey-Bass, San Francisco, CA,
US, pp. 349–381.
Bohan, P., Laing, M., 2012. Can leadership behaviour affect quality and safety? Br. J.
Health Care Manag. 18 (4), 184–190.
Brislin, R., 1980. Translation and content analysis of oral and written material. In:
Triandis, H.C., Berry, J.W. (Eds.), Handbook of Cross-Cultural Psychology, vol. 2.
Allyn & Bacon, Boston, pp. 389–444.
Brondino, M., Silva, S.A., Pasini, M., 2012. Multilevel approach to organizational
and group safety climate and safety performance: co-workers as the missing
link. Saf. Sci. 50 (9), 1847–1856.
Brown, T.A., 2015. Confirmatory factor analysis for applied research, 2nd ed.
Guilford Publications, New York.
Browne, M.W., Cudeck, R., 1993. Alternate ways of assessing model fit. In: Bollen,
K.A., Long, J.S. (Eds.), Testing Structural Equation Models. Newbury Park, CA,
Sage, pp. 136–162.
Cavazza, N., Serpe, A., 2009. Effects of safety climate on safety norm violations:
exploring the mediating role of attitudinal ambivalence toward personal
protective equipment. J. Safety Res. 40 (4), 277–283.
Chen, I.C., Ng, H.F., Li, H.H., 2012. A multilevel model of patient safety culture:
cross-level relationship between organizational culture and patient safety
behavior in Taiwan’s hospitals. Int. J. Health Plann. Manage. 27 (1), 65–82.
Choundry, R.M., Fang, D., Mohamed, S., 2007. The nature of safety culture: a survey
of the state-of-the-art. Saf. Sci. 45, 993–1012.
Christian, M.S., Bradley, J.C., Wallace, J.C., Burke, M.J., 2009. Workplace safety: a
meta-analysis of the roles of person and situation factors. J. Appl. Psychol. 9,
1103–1127.
Clarke, S., 2006. The relationship between safety climate and safety performance: a
meta-analytic review. J. Occup. Health Psychol. 11 (4), 315–327.
Clarke, S., 2013. Safety leadership: a meta-analytic review of transformational and
transactional leadership styles as antecedents of safety behaviours. J. Occup.
Organ. Psychol. 86, 22–49.
Cooper, G.L., Cartwright, S., Earley, P.C., 2001. The International Handbook of
Organizational Culture and Climate. Wiley, New York.
Cooper, M.D.1, 2000. Towards a model of safety culture. Saf. Sci. 36 (2), 111–136.
Cox, S., Flin, R., 1998. Safety culture: philosopher’s stone or man of straw? work &
stress: an international journal of work. Health Organ. 12 (3), 189–201.
DeJoy, D.M., 2005. Behavior change versus culture change: divergent approaches
to managing workplace safety. Saf. Sci. 43 (2), 105–129.
 L. Petitta et al. / Accident Analysis and Prevention 99 (2017) 77–89
EU-OSHA, 2002. Inventory of Socioeconomic Costs of Work Accidents. European
Agency for Safety and Health at work https://osha.europa.eu/en/publications/
reports/207.
Eldridge, J.E.T., Crombie, A., 1974. A Sociology of Organizations. Allen & Unwin,
London.
Enders, C.K., Tofighi, D., 1998. Centering predictor variables in cross-sectional
multilevel models: a new look at an old issue. Psychol. Methods 12 (2),
121–138.
Enriquez, E., 1970. I tipi di strutture decisionali [The types of decisional structures].
In: AA. VV., Les problemes de gestion des enterprises. Paris: Dunod.
European Commission, 2011. Socio-Economic Costs of Accidents at Work and
Work-related Ill Health: Key Messages and Case Studies. European Union,
Luxembourg.
Fang, D.P., Chen, Y., Louisa, W.1, 2006. Safety climate in construction industry: a
case study in Hong Kong. J. Constr. Eng. Manag. 132 (6), 573–584.
Fugas, C.S., Silva, S.A., Melià, J.L., 2012. Another look at safety climate and safety
behavior: deepening the cognitive and social mediator mechanisms. Accid.
Anal. Prev. 45, 468–477.
Glendon, A.I., Stanton, N.A., 2000. Perspectives on safety culture. Saf. Sci. 34,
193–214.
Griffin, M.A., Neal, A., 2000. Perceptions of safety at work: a framework for linking
safety climate to safety performance, knowledge, and motivation. J. Occup.
Health Psychol. 5 (3), 347–358.
Guldenmund, F.W., 2007. The use of questionnaires in safety culture research-an
evaluation. Saf. Sci. 45, 723–743.
Hale, A.R., 2000. Culture’s confusions. Saf. Sci. 34 (1–3), 1–14.
Hartmann, C., Meterko, M., Rosen, A., Zhao, S., Singer, S., Guba, D., 2009.
Relationship of hospital organizational culture and patient safety climate in
the Veterans Health Administration. Med. Care Res. Rev. 66, 320–338.
Heck, R.H., Thomas, S.L., 2015. An Introduction to Multilevel Modeling Techniques:
MLM and SEM Approaches Using Mplus. Routledge/Taylor & Francis, New York.
Hofmann, D.A., Gavin, M.B., 1998. Centering decisions in hierarchical lineal
modeling: implications for research in organizations. J. Manag. 24 (5), 623–641.
Hofmann, D.A., Morgeson, F.P., 2004. The role of leadership in safety. In: Barling,
J.E., Frone, M.R. (Eds.), The Psychology of Workplace Safety. American
Psychological Association, Washington, DC,US, pp. 159–180.
House, R.J., Hanges, P.J., Javidan, M., Dorfman, P.W., Gupta, V., 2004. Leadership,
Culture, and Organizations: The GLOBE Study of 62 Societies. Sage Publications.
International Labour Office, 2006. World of Work Magazine No. 57, September
2006. ILO, Geneva, Switzerland.
Istituto Nazionale per l’Assicurazione contro gli Infortuni sul Lavoro [National
Institute for Insurance against Accidents at Work], 2011. Rapporto Annuale
2014 [Annual Report 2014], Retrieved from: https://www.inail.it/cs/internet/
comunicazione/sala-stampa/infografiche/ucm 184882 relazione-annuale-
2014—infortuni-sul-lavoro-e.html?idPage=2443085357346.
James, L.R., 1982. Aggregation bias in estimates of perceptual agreement. J. Appl.
Psychol. 67, 675–689.
Jansen, M.W., De Leeuw, E., Hoeijmakers, M., De Vries, N.K., 2012. Working at the
nexus between public health policy, practice and research. Dynamics of
knowledge sharing in the Netherlands. Health Res. Policy Syst. 10 (33), 1–9.
Kapp, E.A., 2012. The influence of supervisor leadership practices and perceived
group safety climate on employee safety performance. Saf. Sci. 50, 1119–1124.
Keyzer, M., 2002. Labeling and the realization of cultural values. Economist 150 (4),
487–511.
Kim, E.S., 2012. Technocratic precautionary principle: Korean risk governance of
mad cow disease. J. Risk Res. 15 (9), 1075–1100.
LeBreton, J.M., Senter, J.L., 2008. Answers to 20 questions about interrater
reliability and interrater agreement. Organ. Res. Methods 11 (4), 815–852.
Lloyd, M., Carson, A., 2005. Culture shift: carer empowerment and cooperative
inquiry. J. Psychiatr. Ment. Health Nurs. 12 (2), 187–191.
Marradi, A., 1990. Classification, typology, taxonomy. Qual. Quan. XXIV 2, 129–157,
Available at: http://web.archive.org/web/20031011003312/http://www.unibo.
edu.ar/marradi/classqq.
Martínez-Córcoles, M., Gracia, F., Tomás, I., Peiró, J.M., 2011. Leadership and
employees’ perceived safety behaviors in a nuclear power plant: a structural
equation model. Saf. Sci. 49 (8–9), 1118–1129.
Mearns, K.J., Flin, R., 1999. Assessing the state of organizational safety—culture or
climate? Curr. Psychol. 18 (1), 5–17.
Mearns, K., Kirwan, B., Reader, T.W., Jackson, J., Kennedy, R., Gordon, R., 2013.
Development of a methodology for understanding and enhancing safety in Air
Traffic Management. Saf. Sci. 53, 123–133.
Muthén, L., Muthén, B., 1998–2012. Mplususer’s guide. Version 7. Los Angeles:
Muthén & Muthén.
Muthen, B., Satorra, A., 1995. Complex sample data in structural equation
modeling. Sociol. Methodol. 2, 267–316.
Nahrgang, J.D., Morgeson, F.P., Hofmann, D.A., 2011. Safety at work: a
Meta-analytic investigation of the link between job demands, job resources,
burnout, engagement, and safety outcomes. J. Appl. Psychol. 96, 71–94.
89
Neal, A., Griffin, M.A., 2006. A study of the lagged relationship among safety
climate, safety motivation, safety behavior, and accidents at the individual and
group levels. J. Appl. Psychol. 91 (4), 946–953.
Neal, A., Griffin, M.A., Hart, P.M., 2000. The impact of organizational climate on
safety climate and individual behavior. Saf. Sci. 34 (1–3), 99–109.
Niskanen, T., 1994. Safety climate in the road administration. Saf. Sci. 17 (4),
237–255.
Payne, R., 2000. Climate and culture: how close can they get? In: Ashkanasy, N.M.,
Wilderom, C.P.M., Peterson, M.P. (Eds.), Handbook of Organizational Culture
Climate. Sage, Thousand Oaks, CA.
Petitta, L., Barbaranelli, C., Probst, T., 2014. Intensity & Strength Organizational
Culture Questionnaire. Manual. Hogrefe Italia, Firenze.
Petitta, L., Probst, T., Barbaranelli, C. (in press) Safety culture, moral disengagement,
and accident underreporting. J. Bus. Ethics. 10.1007/s10551-015-2694-1.
Phuong-Mai, N., Terlouw, C., Pilot, A., 2005. Cooperative learning vs confucian
heritage culture’s collectivism: confrontation to reveal some cultural conflicts
and mismatch. Asia Eur. J. 3 (3), 403–418.
Probst, T.M., Brubaker, T.L., 2001. The effects of job insecurity on employee safety
outcomes: cross-sectional and longitudinal explorations. J. Occup. Health
Psychol. 6 (2), 139–159.
Probst, T.M., Graso, M., Estrada, A.X., Greer, S., 2013. Consideration of future safety
consequences: a new predictor of employee safety. Accid. Anal. Prev. 55,
124–134.
Probst, T.M., 2015. Organizational safety climate and supervisor safety
enforcement: multi-level explorations of the causes of accident
under-reporting. J. Appl. Psychol. 100 (6), 1899–1907.
Raudenbush, S.W., Bryk, A.S., 2002. Hierarchical Lineal Modeling: Applications and
Data Analysis Methods, second editions. Sage, Newbury Park, CA.
Raykov, T., Marcoulides, G.A., 2011. Introduction to Psychometric Theory. New
York, NY, US, Routledge/Taylor & Francis Group.
Richter, A., Koch, C., 2004. Integration: differentiation and ambiguity in safety
cultures. Saf. Sci. 42, 703–722.
Ross, J.1, 2011. Patient safety outcomes: the importance of understanding the
organizational culture and safety climate. J. Perianesth. Nurs. 26 (5), 347–348.
Schein, E.H., 1985. Organizational Culture and Leadership. Jossey-Bass, San
Francisco.
Schein, E.H., 1999. Process Consultation Revisited. Addison-Wesley, Reading.
Shrout, P.E., Fleiss, J.L., 1979. Intraclass correlations: uses in assessing rater
reliability. Psychol. Bull. 86 (2), 420–428.
Simard, M., Marchand, A., 1997. Workgroups’ propensity to comply with safety
rules: the influence of micro-macro organisational factors. Ergonomics 40 (2),
172–188.
Spector, P.E., 2006. Method variance in organizational research. Truth or urban
legend? Organ. Res. Method 9 (2), 221–232.
Speroff, T., Nwosu, S., Greevy, R., Weinger, M.B., Talbot, T.R., Wall, R.J., Deshpande,
J.K., France, D.J., Ely, E.W., Burgess, H., Englebright, J., Williams, M.V., Dittus,
R.S., 2012. Organisational culture: variation across hospitals and connection to
patient safety climate. Qual. Saf. Health Care 19 (6), 592–596.
Su, S.F., Jenkins, M., Liu, P.E., 2012. Nurses” perceptions of leadership style in
hospitals: a grounded theory study. J. Clin. Nurs. 21 (1-2), 272–280.
Valsiner, J., Connolly, K. (Eds.), 2002. Sage, London.
Van De Vliert, E., Einarsen, S., 2008. Cultural construals of destructive versus
constructive leadership in major world niches. Int. J. Cross Cult. Manag. 8 (3),
275–295.
Wallace, J.C., Edwards, B., Paul, J., Burke, M., Christian, M., Eissa, G., 2016. Change
the referent? A meta-analytic investigation of direct and referent-shift
consensus models for organizational climate. J. Manag. 42 (4), 838–861.
Yule, S., 2003. Senior Management Influence on Safety Performance in the UK and
US Energy Sectors. Doctoral Thesis. University of Aberdeen, Scotland.
Zohar, D., Luria, G., 2003. The use of supervisory practices as leverage to improve
safety behavior: a cross-level intervention model. J. Safety Res. 34, 567–577.
Zohar, D., Luria, G., 2005. A multilevel model of safety climate: cross-level
relationships between organization and group-level climates. J. Appl. Psychol.
90, 616–628.
Zohar, D., 1980. Safety climate in industrial organizations: theoretical and applied
implications. J. Appl. Psychol. 65 (1), 96–102.
Zohar, D., 2000. A group-level model of safety climate: testing the effect of group
climate on micro-accidents in manufacturing jobs. J. Appl. Psychol. 85,
587–596.
Zohar, D., 2002. The effects of leadership dimensions, safety climate: and assigned
priorities on minor injuries in work groups. J. Organ. Behav. 23, 75–92.
Zohar, D., 2003. Safety climate: conceptual and measurement issues. In: Quick, J.,
Tetrick, L.E. (Eds.), Handbook of Occupational Health Psychology. American
Psychological Association, Washington, DCUS, pp. 123–142.