Você está na página 1de 7

Journal of Safety Research 40 (2009) 277283

Contents lists available at ScienceDirect

Journal of Safety Research


j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / j s r

Effects of safety climate on safety norm violations: exploring the mediating role of
attitudinal ambivalence toward personal protective equipment
Nicoletta Cavazza , Alessandra Serpe 1
Universit di Modena-Reggio Emilia, via Allegri 9, 42100 Reggio Emilia

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

Available online 4 August 2009 Problem: Research on the role of organizational and psychosocial factors in inuencing risk behaviors and the
likelihood of injury at work showed that safety climate also has great impact on workers behavior. However,
Keywords: the mechanisms through which this impact operates are still partially unclear. Method: In order to explore
safety climate the role that attitudinal ambivalence toward wearing PPE might play in mediating the impact of safety
ambivalence
climate on safety norm violations, a questionnaire was administered to 345 Italian workers. Results: Three
safety norm violations
dimensions of safety climate (i.e., company safety concern, senior managers safety concern, supervisors
work injuries prevention
personal protective equipment
attitudes towards safety) were found to be positively associated with the individual ambivalence level,
whereas the fourth one (i.e., work pressure) was negatively correlated with it. In turn, low levels of
ambivalence were associated with a lower tendency to break the safety norms, even though the perception
of a good safety climate also maintained a direct effect on unsafe behaviors. Impact on industry: Designers of
training program for the prevention of work related injuries must pay great attention to the psycho-social factors
(such as the effects of the safety climate perception by employees on their attitudes and behaviors), and include
specic contents into the prevention programs in order to improve workers compliance with safety norms.
2009 National Safety Council and Elsevier Ltd. All rights reserved.

1. Problem possible. Monodisciplinary clinical and biomedical approaches are


now quite well-developed; however, ndings of technological
Statistics provided by the International Labour Organization (ILO) research aiming at the design of safe products and processes are
show that, despite scientic and technological progress, working also now available. The joint effects of features of people, tools, tasks,
conditions in the countries of the European Union have not changed and operating environments on the likelihood of injury remains a less
to such a degree as to signicantly reduce the problem of occupational well-explored eld in Italy (Dekker, 2002). A deeper knowledge from
injuries. While workers pay the highest price of all, there are also this perspective, however, would allow organizations to consciously
social and economic costs, which average about 3% of the gross address the dangers inherent in the production process by promoting
national product of each country in the European Union. Safety safety practices and managing potential residual risk (which cannot
conditions in the workplace vary notably from country to country, be eliminated) by promoting the use of personal protective equipment.
from sector to sector, and from social group to social group (Dembe, The present study examines the potential mediating effect of
1999). The situation in Italy is just as serious, where about 900,000 employees attitudinal ambivalence toward safety practices on the
injuries and over 1,300 deaths were recorded on average per year relationship between organizational safety climate and unsafe
between 1998 and 2005 (data source: Inail). In 2006, the standardized behaviors. Thus, unsafe behaviors are seen as a symptom of something
rate of incidence of fatal work injuries in Italy was 2.9, which is dysfunctional at an organizational level. Greater awareness of how
slightly higher than the European Union (Euro area rate of 2.5; data this relationship works is of particular interest as ndings could
source, Eurostat). This is why research in this eld carried out in Italy potentially orient interventions (e.g., training programs), which could
is both urgent and of interest. Research on occupational safety has result in improved workplace safety.
developed considerably in many disciplines, with the main objective
of eliminating from the workplace as many dangers and risks as 1.1. Safety Climate

We wish to express our gratitude to Anna Rita Graziani, Franco Fraccaroli, Leopoldo Research on the role of organizational and psychosocial factors in
Magelli, Michele Roccato, and Venere Pavone for comments on earlier drafts of this inuencing risk behaviors and the likelihood of injury at work showed
manuscript. that safety climate has great impact (Clarke, 2006).
Corresponding author. Tel.: +39 0522 523262; fax: +39 0522 523055.
E-mail addresses: nicoletta.cavazza@unimore.it (N. Cavazza), aserpe@unimo.it
The construct of a safety climate was introduced by Zohar (1980)
(A. Serpe). and was dened as the employees shared perceptions of safety policies,
1
Tel.: + 39 0522 523262; fax: + 39 0522 523055. procedures, and practices, as well as the overall importance and the true

0022-4375/$ see front matter 2009 National Safety Council and Elsevier Ltd. All rights reserved.
doi:10.1016/j.jsr.2009.06.002
278 N. Cavazza, A. Serpe / Journal of Safety Research 40 (2009) 277283

priority of safety at work (Grifn & Neal, 2000; Zohar, 1980, 2000, 2003). (Brown, Willis, & Prussia, 2000; Cabrera & Isla, 1998; DeJoy, Murphy, &
Safety climate is a specic form of organizational atmosphere that taps Gershon, 1995; Hofmann & Stetzer, 1996; Tomas, Melia, & Oliver, 1999;
into individual perceptions of the importance of safety in the work Varonen & Mattila, 2000), and in the number of injuries (Barling,
environment (Neal, Grifn, & Hart, 2000). It acts as a frame of reference Loughlin, & Kelloway, 2002; Cohen, 1977; Dedobbeleer & Beland, 1998;
that orients employee and workgroup practices and a range of work- Gillen, Baltz, Gassel, Kirsh, & Vaccaro, 2002; Hofmann & Stetzer, 1996;
related individual attitudes (Morrow & Crum, 1998). Mearns, Flin, Fleming, & Gordon, 1998; Mearns, Flin, Gordon, & Fleming,
While there is general agreement on the denition of safety 2001; Zohar, 2000). The meta-analytic review by Clarke (2006) found
climate, the dimensionality of the construct remains controversial strong support for the link between safety climate and employees safety
(Johnson, 2007). Some researchers have argued for a uni-dimensional performance, and a valid, though weak, correlation between safety
latent variable (Neal et al., 2000), while others have claimed that the performance and occupational accidents and injuries.
safety climate is multi-dimensional (e.g., Cooper & Phillips, 2004; Huang et al. (2006) found that the perception of control completely
Dedobbeleer & Beland, 1998; Mohammed, 2002; O'Toole, 2002; mediated the effect of the safety climate on the frequency of occupational
Zohar, 1980; Zohar & Luria, 2005), although they do not agree on the injuries: the safety climate (measured in terms of management
number of factors that constitute it. commitment to safety, return to work policies, post injury administra-
Despite the differences in the number of factors, a recurrent theme tion and safety training) enhanced the workers perception of their
has emerged. Indeed, many dimensions (e.g. perceived management personal ability to avoid injury and accidents and led to a reduction in the
attitudes and behaviors in relation to safety, management concern for number of self-reported work-related injuries. Recently, a longitudinal
employees well-being, perceived importance of safety training survey by Pousette et al. (2008) found that safety climate signicantly
programs) concern management commitment to safety (Cohen & predicted self-reported safety behaviors seven months later.
Cleveland, 1983; Flin, Mearns, Fleming, & Gordon, 1996; Huang, Ho, Most of the research focusing on the relationship between safety
Smith, & Chen, 2006; Mason & Simpson, 1995; Seo, Torabi, Blair, & climate and safety behaviors has shown that the relationship is
Ellis, 2004). Some studies showed that the dimensions related to mediated by other variables (Cooper & Phillips, 2004). The study of
management commitment to safety account for over half of the total mediational factors is important because it allows us to understand
variance in safety climate (Huang et al., 2006; Zohar, 1980). However, the mechanisms through which safety climate operates on those
a number of studies have included different or several levels of workers behaviors that reduce risk of injury.
management (e.g., senior managers, supervisor). Two of the mediational factors that convey the inuence of safety
Cheyne, Cox, Oliver, and Tomas (1998) reported management climate on the likelihood of using personal protective equipment are:
commitment as a primary factor in their model of safety behaviors, the knowledge of the safety procedures and practices in the workplace
and they also incorporated safety communication as one component and the motivation to perform safety-related actions or to participate
of safety climate. Safety communication concerns the perception of in activities supporting safety in the organization (Grifn & Neal,
being able and encouraged to discuss safety issues in the organization 2000). Other studies showed that workers psychological stress (Siu,
and of having access to relevant safety information (Larsson, Pursette, Phillips, & Leung, 2004) and production pressures (Brown et al., 2000)
& Torner, 2008). In fact, a positive safety climate is more likely to exist can play the same role.
in an environment where there is an open and effective exchange of Seo (2005) found that perceived safety climate affected safety
information (DeJoy, Schaffer, Wilson, Vanderberg, & Butt, 2004). performance in three different paths simultaneously: (a) indirect,
Other dimensions regarding safety climate are workgroup safety through the inuence of perceived work pressure affecting perceived
involvement and the priority of safety issues (Pousette, Larsson, & risk inuencing perceived barriers (i.e., skepticism regarding the
Trner, 2008). importance and efcacy of safety procedures); (b) through direct
Workgroup safety involvement concerns workers perception of inuence on perceived barriers that affected unsafe practices; and (c)
coworkers involvement in and commitment to safety issues: if a through direct inuence on unsafe behaviors.
worker perceives that his/her coworkers are concerned about safety, In the present study we focused on six aspects of safety climate
the group tends to practice safety behaviors (Hayes, Perandan, and we explored their role in reducing workers attitudinal ambiva-
Smecko, & Trask, 1998). lence toward using personal protective equipments (hereinafter PPE),
The concept of safety climate must also include all those aspects hypothesizing that the latter has a mediational role between safety
that refer to how management deals with safety issues in relation to climate and safety behaviors (see below). Dimensions of safety
other company objectives (Cox & Cheyne, 2000; Niskanen, 1994; climate were chosen empirically on the basis of both the importance
Tharaldsen, Olsen, & Rundmo, 2008). Safety issues often compete with attributed to them by researchers as illustrated above and the
other operational issues such as speed or ow production. Safety relevance to the Italian context. They were: company safety concern
policies and procedures can be inferred from the workers perceptions (CSC), senior managers safety concern (SMSC), supervisors attitude
about the balance maintained between pressure for production and toward safety (SAS), workgroup safety involvement (WSI), work
safety (Zohar, 2008). Indeed, in a number of surveys, aspects relating pressure (WP), safety communication (SC).
to work pressure are included as part of the broad safety climate (Flin,
Mearns, O'Connor, & Bryden, 2000). 1.2. Attitudinal Ambivalence toward Safety Behaviors
The scarce agreement on the number of dimensions converging in
the construct is also due to the little evidence regarding the validity of In the last decade, many social psychologists involved in the study of
scales that measure the safety climate, given that these scales were attitude within a social-cognitive approach have increasingly focused
often developed in relation to specic national and industrial contexts their attention on the feature of ambivalence. They have moved from a
(Seo et al., 2004). denition of attitudes as a univalent construct (either positive or negative
Aside from the dimensions incorporated, the concept of safety valenced toward an object) to a denition in which positive and negative
climate is of interest because of its inuence on workers behavior. A evaluations of a particular object coexist in the same cognitive structure.
number of studies have shown that safety climate is the main antecedent Ambivalence may be held toward goals, events, objects, and/or
of safety performance (participation and compliance to safety rules and behaviors. People's reactions to conicting stimuli (e.g., to carry
procedures) in organizations (Flin et al., 2000; Grifn & Neal, 2000; Seo, condoms) drives attitudinal ambivalence, and personal behaviors are
2005). Indeed, research has revealed growing evidence that perception particularly likely to be the object of attitudinal ambivalence because
of safety climate is associated with an increase in safety practices of the motivational conicts they generate (Conner & Armitage, 2008;
(Cheyne et al., 1998; Zohar, 1980), a decrease in unsafe behaviors Conner et al., 2002). Among behaviors, self-protecting behaviors are
N. Cavazza, A. Serpe / Journal of Safety Research 40 (2009) 277283 279

often such conicting stimuli due to the conict between immediate Table 1
costs and future potential benets. In the same way, risky behaviors Research participants for each type of industry.

are conicting stimuli due to the conict between immediate benets Type of Number of Response Mean Number of Mean job
and future potential costs. industry participants rate age women tenure
More generally, many health-related behaviors are likely to be Mechanic 288 94% 37.07 6 5.99
associated with ambivalence, since the promotion of such behaviors Textile 38 91% 41.70 11 10.52
Food 81 87% 41.50 22 11.79
in order to improve health can contrast with the perceived cost of that
Total 345 91% 38.67 39 11.20
behavior or with a positive evaluation of an alternative, health-
threatening, behavior. For example, Conner, Povey, Sparks, James, and
Shepherd (1998) examined 12 health-risk and health-protective
behaviors relevant to students and found that drinking alcohol and the mandatory use of PPE on the job), at least in part, through the
eating a low-fat diet produced the highest degrees of ambivalence. reduction of attitudinal ambivalence toward wearing PPE.
Furthermore, societal norms, social groups, and the attitudes of To test this hypothesis, we compared a full mediation, a partial
signicant others may provide potentially conicting information. mediation, and an independent model.
Thus, attitudinal ambivalence may have interpersonal antecedents Our nal goal was to explore which safety climate dimensions had
such as the perception of normative conict (Priester & Petty, 2001). the heaviest impact on attitudinal ambivalence. As the topic is
But interpersonal and social factors could also reinforce one side of the relatively unexplored, we could not formulate any hypothesis.
conict (e.g., make the individual perceive others expectations for
him/her to behave in a prevention way) and weaken the other, 2. Method
thereby reducing individual ambivalence. This is why we can expect
that a positive safety climate could impact the workers level of 2.1. Data and procedure
ambivalence: Positive safety climate communicates that the benet of
following the safety rules outweighs the costs. The present research is based on a questionnaire administered to
Research has shown that attitudinal ambivalence has some 345 blue-collar workers from three companies located in the north of
important implications for the attitude-behavior relationship. Indeed, Italy, as reported in Table 1. The rst company is a mechanic industry
many studies have revealed a moderation effect of ambivalence, such devoted to the construction and maintenance of networks for electrical
that higher levels of ambivalence are related to lower attitude- energy, water, and gas; the second one is an industry designing and
behavior coherence (Armitage & Conner, 2000; Conner et al., 2002; producing creative knitted fabrics; and the third one is a company of
Jonas, Broemer, & Diehl, 2000). The moderation effects were found beef slaughtering and meat processing.
both for self-reported and observed behavior (Sparks, Harris, & The questionnaire was completed on a voluntary basis during
Lockwood, 2004). Some of the studies showing this effect were work time and with at least one researcher present.
carried out specically on health-related behavior (e.g., Dormandy, Participants were 83.5% (n = 288) male, aged from 17 to 63
Hankins, & Marteau, 2006). (M = 38.67, SD = 9.43), and mean job tenure was 11.20 years
However, as far as we know, there are no published studies (SD = 8.75, range 0-38); 48.7% of the respondents had compulsory
focusing on the role of attitudinal ambivalence in the safety work- education level (i.e., 8 years of schooling).
related domain. The present study examines this issue, which is
important for two reasons. First, it provides insight into how different 2.2. Measures
levels of ambivalence toward self-protecting behaviors (e.g., wearing
PPE) are related with the actual frequency of unsafe behaviors (e.g., 2.2.1. Unsafe behavior
not wearing PPE). Second, it tries to extend our comprehension of the In each of the companies taken into consideration, our respon-
mechanisms through which organizational safety climate inuence dents are obliged by the law to wear some kind of PPE during their
actual worker behavior. main task. Thus, the critical dependent variable employed in this
study was the frequency of violations of formal safety norms
1.3. Overview concerning the use of personal protective equipment (PPE). Indeed,
unsafe behavior is considered a valid proxy for injury and the best
The main goal of our research was to show that attitudinal predictor of accidents/near misses as measured by self-reported data
ambivalence toward safety behaviors is a critical factor worth (Mearns et al., 2001). Moreover, reviews of injury records showed
studying in order to better understand how to reduce injury risks in that the majority of injuries were associated with unsafe acts (Feyer,
the workplace. We intended to rst explore whether in the work Williamson, & Cairns, 1997).
domain, as in other health-related domains, individual ambivalence Among a variety of potential violations, we chose to ask
toward safety behaviors increases the likelihood of performing risky respondents about the frequency of violations in the use of PPE
behaviors. In line with the literature on attitudinal ambivalence, we because, besides the specic practices to prevent production-related
formulated the following hypothesis: risks, wearing some item of PPE is required of workers in every sector.
We included three items focusing on three different PPE in
Hypothesis 1 [H1]. Lower employee ambivalence toward self- accordance with the specic norms regulating safety in each
protective behavior (i.e., to use PPE) will relate with lower frequency company2 (overall Cronbach's Alpha = .66). Participants were
of unsafe behavior (i.e., to work without PPE). requested to indicate the frequency with which they usually do
Our second goal was to determine whether ambivalence could play a not comply with the norms on a 4-point scale ranging from
mediational role between safety climate (including the dimensions of
company safety concern, senior managers safety concern, supervisors
attitude toward safety, workgroup safety involvement, work pressure, 2
The PPE included were following:
safety communication) and risky behaviors.
We formulated the hypothesis as follows: - for the mechanic company: safety glasses, helmet, protective footwear.
- for the textile company:protective footwear, protective gloves, hearing protective
Hypothesis 2 [H2]. Organizational safety climate reduces the devices.
frequency of workers violations of safety rules (i.e., norms regulating - for the food company: thermal protective garment, protective footwear, helmet.
280 N. Cavazza, A. Serpe / Journal of Safety Research 40 (2009) 277283

1 = never to 4 = always. Thus, higher scores indicated more frequent 2.3. Data analysis procedure
unsafe behaviors.
Descriptive statistics and correlations of the studied variables were
2.2.2. Safety climate analyzed rst. Structural equation modeling (SEM) techniques were
In order to measure the workforce's perception of safety climate, then used to test and compare the hypothesized mediational models.
we decided to construct a scale on the basis of the suggestions Amos 4.0 software was used to perform SEM.
advanced by three contributions to the debate regarding how to
measure safety climate (Cheyne et al., 1998; Clarke, 2006; Flin et al., 3. Results
2000). Specically, they acknowledge a number of emergent themes
among a wide range of climate features assessed in the studies that 3.1. Intercorrelations among the variables
mainly concerned management (seniors, plants, supervisors) invol-
vement toward safety, work pressure, and communication about Descriptive statistics and the intercorrelations among the included
safety. Thus, we considered six dimensions, each of which included variables are displayed in Table 2. We generally found that percep-
three items formulated to adapt to the Italian organizational context. tions of company safety concern, senior managers safety concern, and
The rst three dimensions concerned management commitment supervisors attitudes toward safety were positively correlated with
to safety, and addressed different levels of management (company as each other, and negatively correlated with work pressure. The rst
a whole, senior managers, supervisors); the fourth concerns the co- three dimensions were positively correlated both with ambivalence
workers attitude and expectation toward safety behavior; the fth and unsafe behaviors, whereas work pressure was negatively
regards the perception of work pressure; the sixth dimension refers to correlated to them. Thus, when in the direction of a positive safety
safety communication. Denitions and reliability coefcients are the climate, the four dimensions were associated with a low level of
following: individual ambivalence toward the use of PPE, but this was
Company safety concern (CSC): the extent to which the company particularly true for the perceptions of line supervisors positive
is perceived to prioritize safety goals (example of an item: my com- attitude toward safety and the perception of the active senior
pany is concerned about safety-related issues regarding workers) managers support of safety practices.
= .84; As expected (H1), ambivalence was positively correlated with self-
Senior managers safety concern (SMSC): workers perceptions of senior reported frequency of unsafe behaviors.
managers support of safety practices and behavior (example of an item:
senior managers in this company encourage workers to use PPE) =.80; 3.2. Mediating role of attitudinal ambivalence
Supervisors attitudes toward safety (SAS): employees perception of
the importance given to safety by supervisors (example of an item: your To study the mediation effect of attitudinal ambivalence toward
line supervisor believes that workers health is very important) =.68; PPE, we carried out three competing path analyses through a
Workgroups safety involvement (WSI): the degree of workgroup structural equation model (see Fig. 1).
joint involvement in safety (example of an item: your workgroup is Because the two alternative competing models are nested in the
concerned about safety) = .34; full mediation model, we compared the full mediation model with
Work pressure (WP): the workers impression that the company each of the two competing models based on the difference between
emphasizes production goals over safety goals (example of an item: the two 2 values associated with the models, which is also
your line supervisor only wants you to nish your work as quickly as distributed as a 2 distribution.
possible) = .74. This is the only dimension for which a higher score To assess the t, we used a mixture of the recommended t-
means a worse safety climate. indices: The absolute t measures likelihood ratio 2; Root Mean
Safety Communication (SC): perceptions of being able and encour- Square Error of Approximation (RMSEA); the Goodness of Fit Index
aged to discuss safety issues (example of an item: if I have a problem (GFI); the Adjusted Goodness of Fit Index (AGFI); the Normed Fit
concerning safety I can talk about it with the line supervisors) = .53; Index (NFI); and the Relative Fit Index (RFI). GFI and AGFI values of 0.9
Since the Cronbach's Alpha showed the two dimensions Safety or above are taken to indicate a good model t to the sample data
Communication and Workgroups safety involvement as having insufcient (Kelloway, 1998). Values of NFI and RFI indicate the degree to which
reliability, we decided to drop them out in the subsequent analyses. the model ts better than a poorly tting model (Bentler & Bonett,
1980). Values of RMSEA are acceptable if they are equal to or lower
2.2.3. Ambivalence than 0.10 (Steiger, 1990).
We measured potential ambivalence toward using PPE. Four items The latent construct of safety climate was indicated by the four
were used to separately assess positive and negative emotions and beliefs dimensions: company safety concern, senior managers safety con-
in reaction to the use of each PPE (example concerning protective cern, supervisors attitudes toward safety, and work pressure.
footwear: wearing protective footwear makes working more difcult; The results showed all three models as having acceptable t (see
wearing protective footwear makes me avoid a possible damage; wearing Table 3). When comparing models B and C against model A, the 2
protective footwear while I work bothers me; wearing protective
footwear makes me feel safe). Thus, there were a total of 12 items for
individuals to respond to on a 4-point scale ranging from 1=completely
disagree to 4=completely agree. We combined the answers in a global
Table 2
ambivalence score through the formula suggested by Thompson, Zanna, Descriptive statistics and intercorrelations (N = 345).
and Grifn (1995). This formula produces a score that is a function of the
simultaneous intensity of the positive and negative ratings. Accordingly, M SD 1 2 3 4 5 6

we calculated potential ambivalence toward the use of PPE by averaging 1. unsafe behavior 2.16 1.10 - .23 - .23 - .22 - .20 .12
2. ambivalence 1.35 .90 - - .31 - .34 - .35 .25
the positive and negative attitude scores (both expressed in positive
3. CSC 3.10 .63 - .72 .58 - .39
values) and subtracting the absolute difference between the two 4. SMSC 3.10 .67 - .60 - .40
components from the average of the two components, using the formula 5. SAS 3.31 .76 - - .41
(P+N)/2-|P-N|+0.5, where P is the positive attitude score, and N the 6. WP 2.65 .79 -
negative attitude score. A higher score indicates a greater level of Note: p b .01; p b .05. CSC = Company safety concern; SMSC = senior managers
ambivalence (range 0-3.50). safety concern; SAS = supervisors attitude toward safety; WP = work pressure.
N. Cavazza, A. Serpe / Journal of Safety Research 40 (2009) 277283 281

4. Discussion

The present results showed that attitudinal ambivalence toward


wearing PPE is a critical factor in predicting workers tendency to
transgress the norms regulating PPE use. Indeed, three dimensions
of safety climate (i.e., company safety concern, senior managers
safety concern, and supervisors attitudes toward safety) were found
to be positively associated with the individual ambivalence level,
and the fourth one (i.e., work pressure) was negatively associated
with it. In sum, when workers perceive that an organizational atmo-
sphere supports safety, they express a low level of ambivalence
toward using PPE. This was especially true for the perceptions
relating to the more proximal level of management (since we had
to drop the workgroup involvement measure out because of its
insufcient reliability). In turn, low levels of ambivalence were
associated with a lower tendency to break the safety norms, even
though the perception of a good safety climate also showed a direct
effect on unsafe behaviors.
Three limitations should be noted. First, the research design is
retrospective because we asked about the frequency of violations in
the period prior to administration of the safety climate scale. The use
of cross sectional data leads to the obvious restrictions of allowing no
conclusions concerning causality. Furthermore, since our constructs
share a common method of measurement, the intervention of a
possible confounding inuence on our results due to common method
variancetypically at work in behavioral research vulnerable to social
desirabilitycould not be excluded (Podsakoff, MacKenzie, Lee, &
Podsakoff, 2003). A worthwhile direction for future studies in this
domain would be to design longitudinal research in order to further
support the observed relationships among constructs.
Second, the respondents may not be able to correctly report the
frequency with which they have transgressed the norms regulating the
use of PPE. Although self-reported data are generally considered a
primary source of information in psychology and social sciences
(Schwarz, 1999), here we are dealing with formally sanctioned
Fig. 1. Full (A), partial (B), and independent (C) path-analytic models: inuence of
behaviors and this may result in their underestimation. Future studies
safety climate on unsafe behaviors through attitudinal ambivalence. Legend: CSC = could employ observed and self-reported measures of violation/
Company safety concern; SMSC = senior managers safety concern; SAS = supervisors compliance and also different kinds of safety behaviors (e.g., employees
attitude toward safety; WP = work pressure. participation in safety activities). The third limitation concerns the
sample construction: generalizability of the present ndings may be
limited because of a lack of probability sampling and the low number of
participants. However, an adequate sample size when using SEM is
difference test revealed model B as the best representation of the considered to be between 25:1 (25 times the number of estimated
empirical data of the three. Furthermore, all the standardized parameters) and 10:1, with a minimum of 200 participants (Kline,
regression weights were signicant and in the predicted direction. 1998). Our study, with 345 participants and a maximum of 7
In this model, R2 for the endogenous variables were 0.61 for company parameters estimated, allowed a ratio of about 50:1.
safety concern, 0.73 for senior managers safety concern, 0.51 for Notwithstanding these limitations, we believe that the present
supervisors attitudes toward safety, 0.24 for work pressure, 0.17 for research contributes to the organizational safety literature by
ambivalence, and 0.11 for unsafe behavior. providing empirical evidence of a mechanism by which safety climate
Thus, the nal structural model showed safety climate as inuencing can alter attitudinal balance, affecting safety behavior.
self-reported unsafe behavior. This inuence occurred, in line with H2,
through both direct and indirect pathways (i.e., different mechanisms 5. Impact on industry
were at work). Regarding the indirect pathway, the results showed that
safety climate reduced the frequency of self-reported violations of the A deep knowledge of the processes by which safety climate affects
norms regarding the use of PPE through the reduction of the employees safety behavior would allow organizations to understand why risky
ambivalence toward wearing PPE, while the direct pathway remains behaviors are sometimes enacted and would help to target the efforts
signicant after controlling for ambivalence. of the prevention programs.

Table 3
2 statistics and t indices for the three proposed models.

Model Chi-square Df p-value Chi square difference relative model A Df difference GFI AGFI NFI RFI RMSEA
A full 25.17 9 .003 .98 .94 .96 .93 .072
B partial 11.65 8 .17 13.52 1 .99 .97 .98 .96 .036
C independent 19.44 9 .03 5.73 0 .98 .95 .97 .95 .058
282 N. Cavazza, A. Serpe / Journal of Safety Research 40 (2009) 277283

Faced with the need to convince workers to adopt safe behaviors, Dormandy, E., Hankins, R., & Marteau, T. M. (2006). Attitudes and uptake of a screening
test: the moderating role of ambivalence. Psychology and Health, 21, 499511.
companies often limit themselves to providing their employees with Feyer, A. M., Williamson, A. M., & Cairns, D. R. (1997). The involvement of human
information regarding current laws governing these behaviors. The behaviour in occupational accident: Errors in context. Safety Science, 25, 5565.
Italian law that regulates those aspects concerning workplace safety Fishbein, M., & Ajzen, I. (1975). Beliefs, attitude, intentions, and behavior. An introduction
to theory and research. Reading, Mass: Addison-Wesley.
acknowledges the fundamental role of training programs for this Flin, R., Mearns, K., Fleming, M., & Gordon, R. (1996). Risk perception an safety in
specic domain. However, the law appears to conceive professional offshore workers. Safety Science, 22, 131145.
training, rst and foremost, as instruction about how to perform a Flin, R., Mearns, K., O'Connor, R., & Bryden, R. (2000). Measuring safety climate:
Identifying the common features. Safety Science, 34, 177192.
task safely. Instead, it is well known that individuals do not act only on Gillen, M., Baltz, D., Gassel, M., Kirsh, L., & Vaccaro, D. (2002). Perceived safety climate,
the basis of the information in their possession but above all in job demands, and co-worker support among union and non-union injured
relation to the expectations or implicit norms that govern the specic construction workers. Journal of Safety Research, 33, 3351.
Grifn, M. A., & Neal, A. (2000). Perception of safety at work: a framework for linking
social context in which they nd themselves (Fishbein & Ajzen, 1975).
safety climate to safety performance, knowledge, and motivation. Journal of
These implicit norms about the relative importance of safety in Occupational Health Psychology, 5, 347358.
respect to other objectives within a company and the obligations that Hayes, B. E., Perandan, J., Smecko, T., & Trask, J. (1998). Measuring perceptions of
derive from formal norms may contradict each other, thereby workplace safety: Development and validation of the Work Safety Scale. Journal of
Safety Research, 29, 145161.
fostering attitudinal ambivalence in the individual, which weakens Hofmann, D. A., & Stetzer, A. (1996). A cross-level investigation of factors inuencing
the chances that this individual will adopt safe behaviors. The climate unsafe behaviours and accident. Personnel Psychology, 49, 307339.
of safety that one breathes in a company is a concentration of Huang, Y. H., Ho, M., Smith, G. S., & Chen, P. Y. (2006). Safety climate and self-reported
injury: assessing the mediating role of employee safety control. Accident Analysis
expectations and possible conicts between implicit and explicit and Prevention, 38, 425433.
norms. Johnson, S. E. (2007). The predictive validity of safety climate. Journal of Safety Research,
Our ndings suggest that great attention must be paid to the effects 38, 511521.
Jonas, K., Broemer, P., & Diehl, M. (2000). Attitudinal ambivalence. European Review of
of the safety climate on psycho-social factors such as attitudes toward Social Psychology, 11, 3574.
violating the formal safety norms. Thus, researchers should further Kelloway, E. K. (1998). Using LISREL for structure equation modeling. Thousand Oaks:
explore the role of attitudinal ambivalence toward a more general class Sage Publication.
Kline, P. (1998). The new psychometrics: Science, psychology and measurement. Florence:
of unsafe practices and practitioners should include specic contents Taylor & Frances.
focusing on psycho-social factors into the prevention programs in order Larsson, S., Pursette, A., & Torner, M. (2008). Psychological climate and safety in the
to improve workers compliance with safety norms. construction industry-mediated inuence on safety behaviour. Safety Sciences, 46,
405412.
Mason, S., & Simpson, G. (1995). Measuring safety attitude to target management action.
Nov: The Safety and Health Practitioner 1720.
References Mearns, K., Flin, R., Fleming, M., & Gordon, R. (1998). Measuring safety climate on
offshore installation. Work & Stress, 12, 238254.
Armitage, C., & Conner, M. T. (2000). Attitudinal ambivalence: A test of three key Mearns, K., Flin, R., Gordon, R., & Fleming, M. (2001). Human and organizational factors
hypotheses. Personality and Social Psychology Bulletin, 26, 14211432. in offshore safety. Work & Stress, 15, 144160.
Barling, J., Loughlin, C., & Kelloway, E. K. (2002). Development and test a model linking Mohammed, S. (2002). Safety climate in construction site environment. Journal of
safety- specic transformational leadership and occupational safety. Journal of Construction Engineering and management, 128, 375384.
Applied Psychology, 87, 488496. Morrow, P. C., & Crum, M. R. (1998). The effects of perceived and objective safety risk on
Bentler, P. M., & Bonett, D. G. (1980). Signicance tests and goodness of t in the employee outcomes. Journal of Vocational Behavior, 53, 300313.
analysis of covariace structure. Psychological Bulletin, 88, 588606. Neal, A., Grifn, M. A., & Hart, P. M. (2000). The impact of organizational climate on
Brown, K. A., Willis, P. G., & Prussia, G. E. (2000). Predicting safe employee behavior in safety climate and individual behaviour. Safety Science, 34, 99109.
the steel industry: Development and test of socio-technical model. Journal of Niskanen, T. (1994). Safety climate in the road administration. Safety Science, 17,
Operations Management, 18, 445465. 237255.
Cabrera, D. D., & Isla, R. (1998). The role of safety climate in a safety management O'Toole, M. F. (2002). The relationship between employee perception of safety and
system. In A. Hale & M. Baram (Eds.), Safety Management: The challenge in change. organizational culture. Journal of Safety Research, 33, 231243.
Oxford: Elsevier. Podsakoff, P. M., MacKenzie, S. B., Lee, J., & Podsakoff, N. P. (2003). Common variance
Cheyne, A., Cox, S., Oliver, A., & Tomas, J. M. (1998). Modelling safety climate in the biases in behavioral research: A critical review of the literature and recommended
prediction of levels of safety activity. Work and Stress, 12, 255271. remedies. Journal of Applied Psychology, 88, 879903.
Clarke, S. (2006). The Relationship Between Safety Climate and Safety Performance: A Pousette, A., Larsson, S., & Trner, M. (2008). Safety climate cross-validation, strength
Meta-Analytic Review. Journal of Occupational Health Psychology, 11, 315327. and prediction of safety behaviour. Safety Science, 46, 398404.
Cohen, A. (1977). Factors in successful occupational safety programs. Journal of Safety Priester, J., & Petty, R. E. (2001). Extending the bases of subjective attitudinal
Research, 9, 168178. ambivalence: Interpersonal and intrapersonal antecedents of evaluative tension.
Cohen, H. H., & Cleveland, R. J. T. (1983). Safety programs practices in recording holding Journal of Personality and Social Psychology, 80, 1934.
plats. Professional Safety, 28, 2633. Schwarz, N. (1999). Self-reported: How the questions shape the answers. American
Conner, M. T., & Armitage, C. J. (2008). Attitudinal ambivalence. In W. Crano & R. Prislin Psychologist, 54, 93105.
(Eds.), Attitudes and Persuasion (pp. 261286). New York: Psychology Press. Seo, D. C. (2005). An explicative model of unsafe work behavior. Safety Science, 43,
Conner, M. T., Povey, R., Sparks, P., James, R., & Shepherd, R. (1998). Dietary change and 187211.
dietary choice: Contributions from social psychology. In A. Murcott (Ed.), Food Seo, D. C., Torabi, M. R., Blair, E. H., & Ellis, N. T. (2004). A cross-validation of safety
choice: Modern social science denitions and discoveries. London: Longman. climate scale using conrmatory factor analytic approach. Journal of Safety
Conner, M. T., Sparks, P., Povey, R., James, R., Shepherd, R., & Armitage, C. J. (2002). Research, 35, 427445.
Moderator effects of attitudinal ambivalence on attitude-behaviour relationships. Siu, O. L., Phillips, D. R., & Leung, T. W. (2004). Safety climate and safety performance
European Journal of Social Psychology, 32, 705718. among construction workers in Honk Kong. The role of psychological strains as
Cooper, M. D., & Phillips, R. A. (2004). Exploratory analysis of the safety climate and mediators. Accident Analysis and Prevention, 36, 359366.
safety Behaviour relationship. Journal of Safety Research, 35, 497512. Sparks, P., Harris, P. R., & Lockwood, N. (2004). Predictors and predictive effects of
Cox, S. J., & Cheyne, A. J. T. (2000). Assessing safety culture in offshore environments. ambivalence. British Journal of Social Psychology, 43, 371383.
Safety Science, 34, 111129. Steiger, J. H. (1990). Structural model evaluation and modication: An interval
Dedobbeleer, N., & Beland, F. (1998). Is risk perception one of the dimensions of safety estimation approach. Multivariate Behavioral Research, 25, 173180.
climate? In A. Freyer & A. Williamson (Eds.), Occupational Injury: Risk Prevention Tharaldsen, J. E., Olsen, E., & Rundmo, T. (2008). A longitudinal study of safety climate
and Intervention (pp. 7381). London: Taylor & Francis. on the Norwegian continental shelf. Safety Science, 46, 427436.
DeJoy, D. M., Murphy, L. R., & Gershon, R. M. (1995). Safety climate in health care Thompson, M. M., Zanna, M. P., & Grifn, D. W. (1995). Let's not be indifferent about
setting. In A. C. Bittner P.C. Champney (Eds.), Advances in industrial Ergonomics and (attitudinal) ambivalence. In R. E. Petty & J.A. Krosnick (Eds.), Attitude strength:
Safety. London: Taylor and Francis. Antecedents and consequences (pp. 361386). Mahwah, NJ: Erlbaum.
DeJoy, D. M., Schaffer, B. S., Wilson, M. G., Vanderberg, R. J., & Butt, M. M. (2004). Tomas, J. M., Melia, J. L., & Oliver, T. (1999). A cross-validation of a structural equation
Creating safer workplaces: Assessing the determinants and role of safety climate. model of accidents: organizational and psychological variables as predictors of
Journal of Safety Research, 35, 8190. work safety. Work and Stress, 13, 4958.
Dekker, S. W. A. (2002). Reconstructing human contributions to accidents: The new Varonen, U., & Mattila, M. (2000). The safety climate and its relationship to safety
view on error and performance. Journal of Safety Research, 33, 371385. practices, safety of the work environment and occupational accidents in eight
Dembe, A. (1999). Social inequalities in occupational health and health care for work- wood-processing companies. Accident Analysis and Prevention, 32, 761769.
related injuries and illnesses. International Journal of Law and Psychiatry, 22, Zohar, D. (1980). Safety climate in industrial organizations: theoretical and applied
567579. implications. Journal of Applied Psychology, 65, 96102.
N. Cavazza, A. Serpe / Journal of Safety Research 40 (2009) 277283 283

Zohar, D. (2000). A group-level model of safety climate: testing the effect of group Alessandra Serpe has a degree in Psychology from the University of Bologna (Italy).
climate on micro-accidents in manufacturing jobs. Journal of Applied Psychology, 85, She is now a PhD student at the University of Modena-Reggio Emilia (Italy). Her
587596. dissertation topic is about psycho-social factors explaining risky behaviors in
Zohar, D. (2003). Safety climate: conceptual and measurement issues. In J. Q. Campbell organizations.
& L.E. Tetrick (Eds.), Handbook of occupational health psychology. Washington:
American Psychological Association.
Zohar, D. (2008). Safety climate and beyond: A multi-level multi-climate framework.
Safety Science, 46, 376387.
Zohar, D., & Luria, G. (2005). A multilevel model of safety climate: Cross-level
relationships between organization and group-level climates. Journal of Applied
Psychology, 90, 616628.

Nicoletta Cavazza has a PhD in Social Psychology from the University of Bologna
(Italy). She is now an associate professor at the Universit di Modena e Reggio Emilia,
Italy. She teaches Psychology of attitudes and opinions, Political Psychology and
Interview and questionnaire techniques. Her main research interests are psycho-social
factors explaining risky behaviors in organizations, attitude change, attitude
ambivalence, persuasive communication, political psychology, and social aspects of
eating. She is a counselor for local institutions about the evaluation of the educational
programs aimed at preventing work injuries.

Você também pode gostar