Accident Analysis and Prevention 47 (2012) 4551

Contents lists available at SciVerse ScienceDirect

Accident Analysis and Prevention

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

Supervisor vs. employee safety perceptions and association with future injury in
US limited-service restaurant workers
Yueng-Hsiang Huang a, , Santosh K. Verma b,d,e , Wen-Ruey Chang c , Theodore K. Courtney b,d ,
David A. Lombardi b,d , Melanye J. Brennan b , Melissa J. Perry d,f
a

Center for Behavioral Sciences, Liberty Mutual Research Institute for Safety, Hopkinton, MA, USA
Center for Injury Epidemiology, Liberty Mutual Research Institute for Safety, Hopkinton, MA, USA
c
Center for Physical Ergonomics, Liberty Mutual Research Institute for Safety, Hopkinton, MA, USA
d
Department of Environmental Health, Harvard School of Public Health, Boston, MA, USA
e
Department of Family Medicine and Community Health, University of Massachusetts Medical School, Worcester, MA, USA
f
Department of Environmental and Occupational Health, George Washington University, School of Public Health and Health Services, Washington, DC, USA
b

a r t i c l e

i n f o

Article history:
Received 26 April 2011
Received in revised form
15 September 2011
Accepted 23 November 2011
Keywords:
Management commitment to safety
Future injury
Supervisor safety perception
Restaurant workers

a b s t r a c t
Objectives: Many studies have found management commitment to safety to be an important construct
of safety climate. This study examined the association between supervisor and employee (shared and
individual) perceptions of management commitment to safety and the rate of future injuries in limitedservice restaurant workers.
Methods: A total of 453 participants (34 supervisors/managers and 419 employees) from 34 limitedservice restaurants participated in a prospective cohort study. Employees and managers perceptions of
management commitment to safety and demographic variables were collected at the baseline. The survey
questions were made available in three languages: English, Spanish, and Portuguese. For the following
12 weeks, participants reported their injury experience and weekly work hours. A multivariate negative binomial generalized estimating equation model with compound symmetry covariance structure
was used to assess the association between the rate of self-reported injuries and measures of safety
perceptions.
Results: There were no signicant relationships between supervisor and either individual or shared
employee perceptions of management commitment to safety. Only individual employee perceptions
were signicantly associated with future employee injury experience but not supervisor safety perceptions or shared employee perceptions.
Conclusion: Individual employee perception of management commitment to safety is a signicant predictor for future injuries in restaurant environments. A study focusing on
employee perceptions would be more predictive of injury outcomes than supervisor/manager
perceptions.
2012 Elsevier Ltd. All rights reserved.

1. Introduction
Restaurants are one of the largest employers in the United
States. According to the Bureau of Labor Statistics (BLS), restaurants
employed approximately 9.7 million workers, which represented
about 6.4% of the total U.S. workforce in 2007 (Bureau of Labor
Statistics, 2007). The BLS also projects that the number of food
preparation and service workers will increase by 396,000 over

Corresponding author at: Liberty Mutual Research Institute for Safety, 71 Frankland Road, Hopkinton, MA 01748, USA. Tel.: +1 508 497 0208;
fax: +1 508 435 0482.
E-mail address: Yueng-hsiang.Huang@Libertymutual.com (Y.H. Huang).
0001-4575/$ see front matter 2012 Elsevier Ltd. All rights reserved.
doi:10.1016/j.aap.2011.11.023

the 10-year period from 2008 to 2018. This is the fourth largest
projected increase in the number of workers among occupational
categories.
BLS data ranks the restaurant industry third in total count of
injuries and illnesses for industries with 100,000 or more nonfatal cases (after schools and hospitals) (Bureau of Labor Statistics,
2008). The California Workers Compensation Institute reported
that restaurant workers accounted for 6.1% of all California job
injury claims from 2000 to 2008. Total medical and indemnity benet payments on these claims in California amounted to just under
$1.1 billion for the 8-year span (Young, 2010). These statistics illustrate the continuing need to identify ways to reduce workplace
accidents and injuries and to improve overall workplace safety for
restaurant workers.

46

Y.H. Huang et al. / Accident Analysis and Prevention 47 (2012) 4551

1.1. Employee perceptions of management commitment to safety

Safety climate is an organizational factor commonly cited as
an important antecedent of safety in the workplace. Safety climate refers to the workers shared perception of the organizations
policies, procedures, and practices as they relate to the value and
importance of safety within the organization (e.g., Zohar, 1980,
2000, 2002, 2003; Grifn and Neal, 2000). As Zohar suggested,
safety climate is a construct that reects the true priority of safety
within an organization (Zohar, 2000).
The dimension that has been included most often is employee
perception of management commitment to safety. This factor has
been described in terms of management concern for employee
well-being (Brown and Holmes, 1986) and management attitudes
toward safety (Zohar, 1980, 2000; Dedobbeleer and Beland, 1991;
Niskanen, 1994). Other studies characterize management commitment to safety in terms of whether workers perceive that safety is
important to management (Diaz and Cabrera, 1997; Gershon et al.,
2000). Based on theoretical and statistical considerations, Zohar
and Luria (2005) promoted a global factor relating to management
commitment when measuring safety climate. In 2008, Zohar concluded that the core meaning of safety climate concerns managerial
commitment, with all other variables that have been associated
with this construct assuming a secondary role both theoretically
and empirically (Flin et al., 2000; Grifn and Neal, 2000; Zohar and
Luria, 2005; Neal and Grifn, 2006). Although prior research has
examined different dimensions of safety climate (e.g., Donald and
Canter, 1994; Hofmann and Stetzer, 1996, 1998; Siu et al., 2004),
consensus on dimensions other than management commitment to
safety is still lacking.
From the literature, safety climate has been differentiated
into two levels, shared-group level safety climate and individualpsychological safety climate (Christian et al., 2009). Summarizing
from the literature (e.g., James et al., 1978; James and Sells,
1981; Jermier et al., 1989), Christian et al. dened psychological
safety climate as individual perceptions of safety-related policies, practices, and procedures pertaining to safety matters that
affect personal well-being at work. Group-level safety climate
is dened as shared perceptions of the work environment and
characteristics as they pertain to safety matters that affect a
group of individuals (e.g., Grifn and Neal, 2000; Zohar and Luria,
2005). This study rst examined whether employee perceptions
of management commitment to safety, the key dimension of safety
climate, were shared among employees working for the same
restaurant.
A recent literature search found that none of the prior studies on
safety climate specically applied to the restaurant industry. The
current study extends the literature by exploring and examining,
specic to the restaurant industry, the important safety climate
construct of employee perceptions of management commitment to
safety.
1.2. Supervisor vs. employee safety perceptions and injury
outcome
Workers usually develop their individual safety perceptions and
expectations by looking to the actions and safety perceptions of
their supervisors to determine the prioritization and importance
of safety (Zohar, 2000). These perceptions and expectations predict
employees motivation to work safely, which affects employees
safety behaviors and subsequent injury outcomes (Hofmann and
Stetzer, 1996; Mueller et al., 1999; Grifn and Neal, 2000; Zohar,
2002, 2003; Huang et al., 2003). From this line of thought, supervisors and employees might have similar safety perceptions of the
workplace, since they work in the same organization. On the other
hand, supervisors, especially in the restaurant environment, are

usually the ones who implement safety polices and procedures, and
conduct safety training for employees. Employees are the ones who
receive the training and they observe how supervisors implement
these polices and procedures. It is possible that there is a gap or
inconsistency with regard to safety perceptions between the person who implements the safety policies, procedures or programs
and the one who receives these programs and observes these implementations. For example, Gittleman et al. (2010) showed that safety
climate perceptions differed by job level with management perceiving a more positive safety climate compared to workers. Prior
studies of multi-source feedback and performance ratings from different sources have also consistently shown discrepancies among
different groups of raters (e.g., supervisors, subordinates) (Mount,
1984; Harris and Schaubroeck, 1988; Borman, 1997; Conway and
Huffcutt, 1997). These differences may derive from varying perspectives on performance and/or opportunities to observe actual
performance (Lawler, 1967). Therefore, the safety perceptions of
supervisors and their employees might be different for the same
restaurant. This study extends the traditional safety climate literature, which usually focuses on only employees perceptions, and
explores the safety perceptions of supervisors vs. their employees
at the same restaurant.
Purpose 1: To explore whether supervisors perceptions of
management commitment to safety are related to employees perceptions.
From safety climate literature, it has been widely shown that
both individual and shared employee safety climate perceptions
can predict employee safety outcomes, such as safety compliance, safety participation, injury and accident rates (Christian et al.,
2009). The results of prior studies are mainly based on employees
perceptions, not the management teams perceptions. It is suggested, in the performance feedback literature, that in order to
gain a more accurate and in-depth understanding of the safety
status in an organization, feedback should be gathered from multiple sources (e.g., workers, supervisors, and senior management)
(Borman, 1997; Law et al., 2011). It is assumed that multi-source
feedback provides unique information from different perspectives, adding incremental validity to performance evaluation. This
study extends the safety climate literature by examining/exploring
whether supervisors perceptions of management commitment to
safety can predict employee safety outcomes.
Purpose 2: To examine whether both supervisor and employee
perceptions of management commitment to safety can predict future
employee injury experience.

2. Methods
2.1. Survey participants
Several approaches were used to recruit the restaurants for this
study. These included approaching chains, stores or franchisees
that had previously been receptive to research studies by the
investigative team members, approaching restaurant trade associations, direct solicitation of stores or franchisees, and outreach
via the loss control department of a large workers compensation insurance company. Thirty-four limited service restaurants
located in six U.S. states, belonging to three major chains agreed to
participate.
A prospective cohort study was conducted via survey in each of
the 34 restaurants. The nal dataset for the current study included
34 supervisors/managers and 419 employees representing 50% of
the total employees in these restaurants. The response rate for
supervisors was 100%. Only two employees declined to participate in the project (at the time baseline surveys were conducted)
indicating a response rate of 99.6% for employee surveys.

Y.H. Huang et al. / Accident Analysis and Prevention 47 (2012) 4551

2.2. Procedures
Baseline survey data were collected from both supervisors and
employees at the restaurants. After completing the baseline survey,
employees were asked to report weekly, for the following 12 weeks,
their injury experience and the number of hours they worked during the previous week. Participants were given a choice of reporting
their weekly experience by (1) telephone using an interactive voice
response (IVR) system; (2) an internet-based survey; or (3) lling
out written survey forms. The study was part of a larger research
project on restaurant safety and was approved by the appropriate
Institutional Review Boards. Details of the procedure were provided
in Verma et al. (2010).
2.3. Measures
2.3.1. Demographic variables
Demographic variables included in the current study were gender, ethnic background, age, education level, job tenure, which
restaurant chain the employee works for, and number of work
hours per week.
2.3.2. Management commitment to safety
Four survey items adapted from Zohar and Luria (2005) and
Huang et al. (2007) were used to measure the factor of employee
perceptions of management commitment to safety. An example is,
The management team emphasizes safe behavior above all other
activities. Each item used a 5-point Likert scale. Survey data were
collected at baseline. The survey materials were made available
in three languages: English, Spanish, and Portuguese. Translation,
back-translation, and check of meaning consistency among languages, were used when developing the Spanish and Portuguese
survey versions. The Internal Consistency Reliability Coefcient
(Alpha) for the four items was 0.81.

47

1993). There is no strict standard of acceptability of ICC2 values.

Glick (1985) recommended an ICC2 cutoff of 0.60. Schneider
et al. (1998) suggested that a moderate value of ICC2 coupled
with an acceptable Rwg score is sufcient grounds for aggregation. The average ICC2 value for their study was 0.47. LeBreton
and Senter (2008) suggested that, depending on the quality of
the measures being used in the multi-level analysis, researchers
will probably want to choose values between 0.70 and 0.85 to
justify aggregation.
(3) RwgJ: The RwgJ is an assessment of within-group interrater
agreement (James et al., 1993). A median of RwgJ larger than
0.70 was used as the criteria.
2.4.2. Multi-level analyses
Restaurants recruited in the study were clustered within chains
and workers were clustered within restaurants. Multi-level analyses were conducted to examine the relationships. To account for
clustering of participants within restaurants, a multivariate negative binomial generalized estimating equation (GEE) model with
compound symmetry covariance structure (Liang and Zeger, 1986;
Zeger and Liang, 1986) was used to assess the association between
the rate of self-reported injuries, and measures of safety perception. Supervisor perceptions were level 2 (group level) variables,
employee shared perceptions within the same restaurant were
level 2 (group level) variables, and individual employee perceptions were level 1 (individual level) variables. As there were only
three chains, two dummy variables for chains were included in the
regression model to account for clustering of restaurants within
chains. Factors not statistically signicant at the 0.05 level in the
univariate model were not included in the multivariate model. All
statistical analyses were conducted using the SAS system version
9.1 (SAS Institute, Inc., Cary, NC).
3. Results

2.3.3. Future injury rate

Every week for the 12 weeks following the baseline survey,
employees reported their injury experience and the number of
hours worked during the previous week by items How many hours
did you work in the restaurant in the last week? and In the last
week, did you get injured while at work? The overall injury rate for
the 12 weeks was calculated as the total number of injuries divided
by total number of hours worked.
2.4. Data analysis procedure
2.4.1. Homogeneity tests of shared employee perceptions of
management commitment to safety
Homogeneity tests were examined by calculating Intraclass Correlation Coefcient 1 and 2 (ICC1, ICC2) and RwgJ (Bartko, 1976;
James et al., 1984, 1993; Bliese, 2000) to check whether employees
perceptions at the same restaurant (within-group) were consistent
and had agreement. The criteria to determine consensus were as
follows:
(1) ICC1: The ICC1 indicates the extent to which individuals within
the same organization assign the same psychological meaning
to, or agree in their perceptions of, an organizational characteristic (Ostroff and Schmitt, 1993). There are no denitive
guidelines on acceptable ICC1 values. In past research, ICC1 values have ranged from 0 to 0.50, with a median of 0.12 (James,
1982).
(2) ICC2: The ICC2 assesses the relative status of between and
within variability using the average ratings of respondents
within each unit (Bartko, 1976). It indicates reliability at the
aggregate level, or the reliability of means (Ostroff and Schmitt,

Table 1 displays detailed information about the demographic

characteristics of the participating employees and supervisors. In
summary, 65.4% of the 419 participating employees were female;
53.9% classied themselves as White, 21% as Black, and 15.5% as Hispanic. Of the survey versions, 89.5% of the participants chose the
English, 7.64% the Spanish, and 2.86% the Portuguese. The mean
age of employees was 29.8 years (range 1578 years), 34.8% had
not completed high school while 40.6% were high school graduates. Participants averaged 32.8 work hours per week and had been
working at the same restaurant for about 33 months. Among 34
supervisors/managers, 67.6% were female; 79.4% classied themselves as White, 11.8% as Black, and 2.9% as Hispanic. The mean age
of supervisors/managers was 38.4 years (range 2057 years), 35.3%
were high school graduates and another 61.8% attended some college or above. They averaged 49.2 work hours per week and had
been working at the same restaurants for about 50 months.
The average injury rate for employees in the 12 weeks following completion of the initial survey was 5.8 injuries per 2000 work
hours (one full-time equivalent). Burns, cuts, and contusions were
typical. Correlations between study variables from employee surveys are provided in Table 2.
Were employee perceptions of management commitment to
safety shared among employees working for the same restaurant? Results of homogeneity tests showed that ICC1 was 0.097,
ICC2 was 0.59, and median of RwgJ as 0.78. According to Glick
(1985) and Schneider et al. (1998), the results of these homogeneity
tests may provide sufcient grounds for aggregation. As discussed
earlier, LeBreton and Senter (2008) suggested that researchers
will probably want to choose values between 0.70 and 0.85 for
ICC2 to justify aggregation. Since there is no strict standard of

48

Y.H. Huang et al. / Accident Analysis and Prevention 47 (2012) 4551

Table 1
Descriptive information of respondents.
Employees
Job title
Crew member
Cashier
Cook
Others
Total participants
Gender
Male
Female
Total participants
Ethnic background
White
Hispanic
Black
Other
Total participants
Education level
Never attended school
Grade 111
High school graduate
Some college or above
Total participants
Survey version chosen
English
Spanish
Portuguese
Total participants
Age group
1619
2024
2534
3544
4554
5564
65 and older
Total participants

Supervisors

Table 3
Relationship between supervisors and individual employees perceptions of management commitment to safety.
Relative
risk

170 (40.6%)
58 (13.8%)
15 (3.6%)
176 (42.0%)
419

34

145 (34.6%)
274 (65.4%)
419

11 (32.4%)
23 (67.6%)
34

226 (53.9%)
65 (15.5%)
88 (21.0%)
40 (9.6%)
419

27 (79.4%)
1 (2.9%)
4 (11.8%)
2 (5.9%)
34

4 (1.0%)
146 (34.9%)
170 (40.7%)
98 (23.4%)
418 (missing 1)

1 (2.9%)
12 (35.3%)
21 (61.8%)
34

Standard
error

p-Value

DV: Individual employee perception of management commitment to safety

0.1036
0.095
0.284
IV: Supervisor perception of
management commitment to safety
Note. n = 34 for supervisors. n = 419 for employees.

375 (89.50%)
32 (7.64%)
12 (2.86%)
419

34 (100%)

108 (26.2%)
87 (21.1%)
89 (21.6%)
57 (13.8%)
48 (11.7%)
16 (3.9%)
7 (1.7%)
412 (missing 7)

0
3 (8.8%)
9 (26.5%)
12 (35.3%)
8 (23.5%)
2 (5.9%)
0
34

acceptability of ICC2 values, the mean scores of management commitment to safety were rst calculated for each restaurant to
represent the shared perception for multi-level analyses (as a level
2, group level variable) for further analyses.
Regarding Purpose 1: Analyses were conducted to examine the question Do supervisors and employees have similar
safety perceptions? Although the mean score of employee perception of management commitment to safety (mean = 4.0; SD = 0.94;
n = 419) and the mean score of supervisor perception (mean = 4.11;
SD = 0.64; n = 34) were similar, results showed no signicant correlation between supervisor perceptions and employee shared
perceptions (r = 0.17; p > 0.05; n = 34). Furthermore, results from
multi-level analyses, as illustrated in Table 3, showed that
there were no signicant relationships between supervisors and
employees individual perceptions of management commitment to
safety.

Regarding Purpose 2: Data were analyzed to answer the questions of whether supervisor safety perceptions and employee safety
perceptions predicted future employee injury experience. Results
from the univariate analyses in Table 4 showed that the supervisors safety perceptions were not a signicant predictor of future
employee injury rates. Employee shared perceptions of management commitment to safety were also not a signicant predictor.
Only employee individual perceptions of management commitment
to safety were signicantly associated with the rate of injury.
We further investigated whether individual employee perceptions of management commitment to safety could predict future
injury outcomes when controlled for demographic variables. Age,
general weekly work hours (as these two variables were correlated to future injury outcome in Table 2) and restaurant chain
(three restaurant chains were represented by two dummy vectors)
were entered in the equation as control variables. Table 5 results
show that individual employee perceptions of management commitment to safety was still a signicant predictor of future injury
rate after controlling for these demographic variables (Table 5);
the rate ratios were very similar to those found in the univariate
analyses. For each one point increase in employee perceptions of
management commitment to safety, the rate of injury decreased 23%
(95% CI 0.59, 0.99).
4. Discussion
The current study examined, in the restaurant environment,
an important construct emerging from the safety climate literature of employee perceptions of management commitment to
safety. Results showed that there were no signicant relationships between supervisors and employees perceptions. Individual
employee perception of management commitment to safety was a
signicant predictor of future injury for restaurant workers, but
not supervisor safety perceptions or shared employee perceptions.
Interestingly, our results showed that there were no signicant
relationships between supervisors and employees perceptions of
management commitment to safety even within the same restaurant. These results were consistent with studies in the literature
of multi-source feedback and performance ratings from different
sources where there were discrepancies among different groups

Table 2
Intercorrelations among study variables from employee surveys.
Variables
1. Gender
2. Age
3. Education
4. Tenure
5. Work days per week
6. Weekly work hours
7. Perceived management commitment to safety
8. Injury rate in the following 12 weeks

2
0.07

3
0.07
0.20**

0.11
0.50**
0.05

0.08
0.15**
0.07
0.06

7
**

0.13
0.19**
0.26**
0.04
0.54**

0.06
0.01
0.03
0.06
0.07
0.06

8
0.02
0.19**
0.05
0.07
0.07
0.11*
0.11*

Male was coded as 1 and female was coded as 0. Injury rate was calculated as injury frequency controlled by work hours in the following 12 weeks after completing the
surveys.
*
Correlation is signicant at alpha < 0.05, two-tailed.
**
Correlation is signicant at alpha < 0.01, two-tailed.

Y.H. Huang et al. / Accident Analysis and Prevention 47 (2012) 4551

49

Table 4
Univariate analyses of the impacts of safety perceptions on injury outcomes.
Number of participants
DV: Injury rate Employee injury experience in the following 12 weeks (controlled for work hours)
Analysis 1
34
IV: Supervisor perception of management commitment to safety
Analysis 2
419
IV: Individual employee perception of management commitment to safety
Analysis 3
34
IV: Shared employee perception of management commitment to safety
*

Relative risk

95% CI

p-Value

1.38

0.87

2.19

0.17

0.75*

0.58

0.96

0.02

0.71

0.37

1.38

0.32

p < 0.05.

Table 5
Multivariate analyses of whether individual employee perceptions of management commitment to safety predict future injury outcomes above and beyond demographic
variables.
Number of participants
DV: Injury rate Employee injury frequency in the following 12 weeks control for work hours for the 12 weeks
IV:
Age
419
Weekly work hours
419
419
Chain 1 vs. 2
419
Chain 1 vs. 3
419
Individual employee perception of management commitment to safety
*
**

Relative risk

0.57**
0.91
0.86
1.06
0.77*

95% CI

0.44
0.80
0.42
0.56
0.59

p-Value

0.75
1.05
1.79
2.00
0.99

0.00
0.19
0.69
0.86
0.04

p < 0.05.
p < 0.01.

of raters (e.g., supervisors, subordinates) (Harris and Schaubroeck,

1988; Borman, 1997; Conway and Huffcutt, 1997). As suggested
in the literature, these differences may be attributed to varying perspectives on and opportunities to observe performance
(Lawler, 1967). Perhaps the safety perceptions were not consistent between restaurant supervisors and employees because they
may have had different opportunities to observe the safety of
the restaurant as well as varying views on safety performance.
Future qualitative research can further investigate potential explanations for these types of discrepancies. Another possible reason
for this lack of consistency may be due to the study limitation that
we did not use separate scales to distinguish between different
levels of restaurant management (i.e. supervisor level and corporate level). Zohar (2008) proposed that safety climate should be a
multi-level construct: an extension into a multi-level framework
that identies organization-level and group-level safety climates
as distinct constructs with separate measurement scales. As with
most of the prior studies in the safety climate literature, separate
scales were not used in the current study when measuring the
factor of management commitment to safety for different levels of
the management team. There is a possibility that the supervisors
might have considered the senior-level managers of the restaurant chain as the managers when answering the survey, and the
employees might consider their direct supervisors as the managers. In this case, supervisors and employees may have used
different points of reference when they responded to the survey
items. Future studies can reduce the potential conceptual ambiguity for the participants by using separate scales to measure
safety commitment of senior management and those of individual
supervisors.
Can supervisors safety perceptions and employees safety
perceptions of management commitment to safety predict future
employee injury experience? The results showed that only individual employee safety perceptions were predictors of employee
future injury experience (but not shared employee safety perceptions or supervisor safety perceptions). From the performance
feedback literature, it is assumed that different rating sources
provide unique information from different perspectives, adding
incremental validity to performance assessment over the individual sources (Borman, 1997; Law et al., 2011). However, in our study

for the restaurant environment, even though it is consistent with

the literature that supervisors had different safety perceptions from
employees, their perceptions did not link to future injury rates at
the restaurant. This suggests that, when investigating the safety of
restaurants, it would be more useful to measure the perceptions of
employees rather than those of the supervisors.
This result is consistent with the traditional safety climate literature which usually focuses on employees perceptions rather
than managers perceptions. Future studies can further examine whether supervisors perceptions are linked to other types
of safety outcomes (e.g., safety behaviors, safety compliance)
and whether they can provide any incremental validity to safety
performance assessment in addition to employees perceptions. Literature pertaining to multi-source ratings has shown that leaders
who received low ratings from others, but overrated themselves,
reported stronger motivation to change (Atwater and Brett, 2005).
Applying the concept from this line of thought, future research can
further explore whether this result that only individual employees
perceptions on safety (not the managers/supervisors) can predict
safety outcomes, can provide useful feedback to motivate supervisors/managers to more objectively evaluate the safety status of
their companies.
Based on commonly acceptable criteria (Glick, 1985; Schneider
et al., 1998), even through the ICC2 score was moderate in the current study, the results of the homogeneity tests provided sufcient
grounds for aggregation. It could, therefore, be reported that there
was a shared perception of management commitment to safety for
restaurant employees. However, the nding of a non-signicant
relationship between group level safety perception and injury was
a result that is contradictory to the safety climate literature (e.g.,
Christian et al., 2009). We do not know, without further investigation, whether or not the specic characteristics of the restaurant
industry (e.g., high turnover rate, generally not a career job) had
an impact on group level safety climate. Also, the restaurants participating in the study belonged to three large chains and may,
in general, have had good safety climate levels and, therefore,
smaller variances between restaurants to be linked to injury experience. In addition, there were only 34 groups (restaurants) and
self-reports of injury may have introduced non-differential misclassication. These factors may have had an impact on why no

50

Y.H. Huang et al. / Accident Analysis and Prevention 47 (2012) 4551

signicant association was found between shared employee perceptions and rate of injury.
On the other hand, some studies in the literature have suggested
that more stringent criteria should be used when discussing ICCs.
For example, LeBreton and Senter (2008) suggested researchers
should choose values between 0.70 and 0.85 to justify aggregation.
In our case, if more stringent criteria were chosen, the results of
the current study would be that there were no group-level shared
employee perceptions on management commitment to safety; therefore, there was no need to test whether there was a relationship
between group-level perception and future injury. As there are no
strict standards of acceptability for the ICC2 value, the contradictory results of the current study may provide illustration that more
stringent criteria may be more appropriate when judging ICCs. This
debate provides a new direction in terms of the need to explore
what is the appropriate ICC value when conducting multi-level
research.
There were several limitations in the current study. We used
a relatively sensitive case denition of injury. Such a case denition could lead to non-differential misclassication of outcome
and could bias the results to the null. However, despite this potential inuence, individual perception of management commitment
to safety assessed at baseline showed a signicant association
with prospectively collected self-reported injury rates. In addition,
although the current study collected weekly self-reported injury
information for 12 weeks, which has less recall bias than a typical
cross-sectional survey, future studies should try to include objective records of injury incidence when examining safety outcomes.
Some limitations with self-reported data (e.g., social desirability
effects) may occur and the data needs to be interpreted with caution.
As mentioned earlier, future studies can use separate scales to
measure safety commitment of senior management and those of
direct supervisors when collecting data from employees as suggested from the recent literature (Zohar, 2008). Survey data can
also be collected from other sources, in additional to employees
and supervisors (e.g., middle level managers, top level managers,
safety personnel) and further examined as to whether their safety
perceptions add incremental validity to safety performance assessment. Also, our study did not provide actual reasons why employees
and supervisors at the same restaurant might have different safety
perceptions of the same organization, which future qualitative
research could investigate. Finally, restaurants belonging to three
major chains and from six different states participated in the study.
The survey material was made available in three different languages, thus increasing the generalizability of the study ndings.
However, since restaurants were mainly owned by large employers, generalizability of some of the ndings may be limited for small
employers. Caution should also be taken when generalizing these
results to other industries.

5. Conclusion
In conclusion, this study showed that supervisors and employees for the same restaurants did not have similar safety perceptions
on management commitment to safety an important construct
emerging from the safety climate literature. Supervisors safety
perceptions were not linked to safety outcomes. Only individual
employee perceptions were predictors of future employee injury
experience. This suggests that when investigating the safety of
restaurants, it would be more useful to measure the perceptions of
employees rather than those of the supervisors. This study provides
support to the traditional safety climate literature which usually
focuses on employees perceptions rather than managers perceptions.

References
Atwater, L.E., Brett, J.F., 2005. Antecedents and consequences of reactions to developmental 360 feedback. Journal of Vocational Behavior 66 (3), 532548.
Bartko, J.J., 1976. On various intraclass correlation reliability coefcients. Psychological Bulletin 83, 762765.
Bliese, P., 2000. Within-group agreement, non-independence, and reliability. In:
Klein, K., Kozlowski, S. (Eds.), Multi-level Theory, Research, and Methods in
Organizations. Jossey-Bass, San Francisco, CA, pp. 349381.
Borman, W.C., 1997. 360 ratings: an analysis of assumptions and a research agenda
for evaluating their validity. Resource Management Review 7, 299315.
Brown, R.L., Holmes, H., 1986. The use of factor-analytic procedures for assessing the
validity of an employee safety climate model. Accident Analysis and Prevention
18, 455470.
Bureau of Labor Statistics, 2007. Occupational Employment Statistics. United
States Department of Labor, U.S. Government Printing Ofce, Washington, DC
(accessed 2009) http://www.bls.gov/oes/oes dl.htm.
Bureau of Labor Statistics, 2008. Census of Fatal Occupational Injuries (CFOI) Current and Revised Data., http://www.bls.gov/iif/oshcfoi1.htm.
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. Journal of Applied
Psychology 94, 11031127.
Conway, J.M., Huffcutt, A.I., 1997. Psychometric properties of multisource performance ratings: a meta-analysis of subordinate, supervisor, peer, and self-ratings.
Human Performance 10, 331360.
Dedobbeleer, N., Beland, F., 1991. A safety climate measure for construction sites.
Journal of Safety Research 22, 97103.
Diaz, I.R., Cabrera, D.D., 1997. Safety climate and attitude as evaluation measures of
organizational safety. Accident Analysis and Prevention 29, 643650.
Donald, I., Canter, D., 1994. Employee attitudes and safety in the chemical industry.
Journal of Loss Prevention in Process Industries 7 (3), 203208.
Flin, R., Mearns, P., OConnor, R., Bryden, R., 2000. Measuring safety climate: identifying the common features. Safety Science 34, 177192.
Gershon, R.R.M., Karkashian, C.D., Grosch, J.W., Murphy, L., Escamilla-Cejudo, A.,
Flanagan, P.A., Bernacki, E., Kasting, C., Martin, L., 2000. Hospital safety climate
and its relationship with safe work practices and workplace exposure incident.
American Journal of Infection Control 28, 211221.
Gittleman, J.L., Gardner, P.C., Haile, E., Sampson, J.M., Cigularov, K.P., Ermann, E.D.,
Stafford, P., Chen, P.Y., 2010. CityCenter and Cosmopolitan Construction Projects,
Las Vegas, Nevada: lessons learned from the use of multiple sources and mixed
methods in a safety needs assessment [case study]. Journal of Safety Research
41 (3), 263281.
Glick, W.H., 1985. Conceptualizing and measuring organizational and psychological
climate: pitfalls in multilevel research. Academy of Management Review 10,
601616.
Grifn, M.A., Neal, A., 2000. Perceptions of safety at work: a framework for linking
safety climate to safety performance, knowledge, and motivation. Journal of
Occupational Health Psychology 5, 347358.
Harris, M.M., Schaubroeck, J., 1988. A meta-analysis of selfsupervisor, self-peer, and
peer-supervisor ratings. Personnel Psychology 41, 4362.
Hofmann, D.A., Stetzer, A., 1996. A cross-level investigation of factors inuencing
unsafe behaviors and accidents. Personnel Psychology 49 (2), 307339.
Hofmann, D.A., Stetzer, A., 1998. The role of safety climate and communication
in accident interpretation: implications for learning from negative events.
Academy of Management Journal 41 (6), 644646.
Huang, Y.H., Chen, P.Y., Rogers, D.A., Krauss, A.D., 2003. Role of workplace safety:
a revisit of the relationship between job satisfaction and employment status. Work: A Journal of Prevention, Assessment & Rehabilitation 21 (3),
251256.
Huang, Y.H., Leamon, T., Courtney, T., Chen, P.Y., DeArmond, S., 2007. Corporate
nancial decision-makers perceptions of safety. Accident Analysis and Prevention 39, 767775.
James, L.R., 1982. Aggregation bias in estimates of perceptual agreement. Journal of
Applied Psychology 67 (2), 219229.
James, L.R., Demaree, R.G., Wolf, G., 1984. Estimating within-group interrater reliability with and without response bias. Journal of Applied Psychology 69, 8598.
James, L.R., Demaree, R.G., Wolf, G., 1993. rwg : an assessment of within-group interrater agreement. Journal of Applied Psychology 78, 306309.
James, L.R., Hater, J.J., Gent, M.J., Bruni, J.R., 1978. Psychological climate: implications
from cognitive social learning theory and interactional psychology. Personnel
Psychology 31, 783813.
James, L.R., Sells, S.B., 1981. Psychological climate: theoretical perspectives and
empirical research. In: Magnusson, D. (Ed.), Toward a Psychology of Situations:
An Interactional Perspective. Erlbaum, Hillsdale, NJ, pp. 275295.
Jermier, J.M., Gaines, J., McIntosh, N.J., 1989. Reactions to physically dangerous work:
a conceptual and empirical analysis. Journal of Organizational Behavior 10 (1),
1533.
Law, R., Tuckey, M.R., Dormann, C., 2011. Psychosocial safety climate as a lead
indicator of workplace bullying and harassment, job resources, psychological
health and employee engagement. Accident Analysis and Prevention 43 (5),
17821793.
Lawler, E.E., 1967. The multitrait-multirater approach to measuring managerial job
performance. Journal of Applied Psychology 51, 369381.
LeBreton, J.M., Senter, J.L., 2008. Answers to twenty questions about interrater
reliability and interrater agreement. Organizational Research Methods 11,
815852.

Y.H. Huang et al. / Accident Analysis and Prevention 47 (2012) 4551

Liang, K.Y., Zeger, S.L., 1986. Longitudinal data analysis using generalized linear
models. Biometrika 73, 1322.
Mount, M.K., 1984. Psychometric properties of subordinate ratings of managerial
performance. Personnel Psychology 37, 687702.
Mueller, L.M., DaSilva, N., Townsend, J.C., Tetrick, L.E., 1999. An empirical investigation of competing safety climate measurement models. In: Proceedings of
14th Annual Meeting of the Society for Industrial and Organizational Psychology,
Dallas, TX.
Neal, A., Grifn, M.A., 2006. A study of the lagged relationships among safety climate,
safety motivation, safety behavior, and accidents at the individual and group
levels. Journal of Applied Psychology 91, 946953.
Niskanen, T., 1994. Assessing the safety environment in the work organization of
road maintenance jobs. Accident Analysis and Prevention 26, 2739.
Ostroff, C., Schmitt, N., 1993. Congurations of organizational effectiveness and efciency. The Academy of Management Journal 36, 6.
Schneider, B., White, S.S., Paul, M.C., 1998. Linking service climate and customer perceptions of service quality: test of a causal model. Journal of Applied Psychology
83, 150163.
Siu, O., Phillips, D.R., Leung, T., 2004. Safety climate and safety performance among
construction workers Hong Kong: the role of psychological strains as mediators.
Accident Analysis and Prevention 36 (3), 359366.
Verma, S.K., Chang, W., Courtney, T.K., Lombardi, D.A., Huang, Y.H., Brennan, M.J., Perry, M.J., 2010. Workers experience of slipping in US limited

51

service restaurants. Journal of Occupational & Environmental Hygiene 7 (9),

491500.
Young, B., 2010. Press Release March 1, 2010: CWCI Scorecard Examines Restaurant
Claims in the California Workers Comp System. California Workers Compensation Institute, Oakland CA, http://www.cwci.org/press release.html?id=149.
Zeger, S.L., Liang, K.Y., 1986. Longitudinal data analysis for discrete and continuous
outcomes. Biometrics 42 (1), 121130.
Zohar, D., 1980. Safety climate in industrial organizations: theoretical and applied
implications. Journal of Applied Psychology 65, 96102.
Zohar, D., 2000. A group-level model of safety climate: testing the effects of group
climate on microaccidents in manufacturing jobs. Journal of Applied Psychology
85, 587596.
Zohar, D., 2002. Modifying supervisory practices to improve subunit safety: a
leadership-based intervention model. Journal of Applied Psychology 87 (1),
156163.
Zohar, D., 2003. Safety climate: conceptual and measurement issues. In: Quick, J.C.,
Tetrick, L.E. (Eds.), Handbook of Occupational Health Psychology. American Psychological Association, Washington, DC, pp. 123142.
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.