Safety climate, safety behavior, and worker injuries in ...

60 downloads 102138 Views 333KB Size Report
ers from 42 companies in Zhongshan City, China. A structured ... co-worker support (Seo et al., 2004; Olsen, 2010; O Connor et al.,. 2011 ... (W. Chen). Safety Science 78 (2015) 173–178 .... The model fit was considered to be good if the.
Safety Science 78 (2015) 173–178

Contents lists available at ScienceDirect

Safety Science journal homepage: www.elsevier.com/locate/ssci

Safety climate, safety behavior, and worker injuries in the Chinese manufacturing industry Xinxia Liu a,b, Guoxian Huang b, Huiqiang Huang b, Shuyu Wang b, Yani Xiao a, Weiqing Chen a,⇑ a b

Faculty of Medical Statistics and Epidemiology, School of Public Health, Sun Yat-sen University, Guangzhou, China Center for Disease Control and Prevention, Zhongshan, Guangdong, China

a r t i c l e

i n f o

a b s t r a c t

Article history: Received 8 October 2014 Received in revised form 31 March 2015 Accepted 27 April 2015

It is estimated that over 10,000,000 occupational injuries occur in China each year. This study explored the relationships between four dimensions of safety climate (management commitment, safety supervision, coworker support, and safety training), three dimensions of safety behavior (safety compliance, personal protective equipment, and safety initiatives), and occupational injuries among Chinese manufacturing workers. A cross-sectional survey was conducted using a sample of 3970 manufacturing workers from 42 companies in Zhongshan City, China. A structured questionnaire was used to capture participants’ socio-demographic characteristics, occupational safety climate, occupational safety behavior, and occupational injuries in the previous year. Path analysis was used to test the relationships between safety climate, safety behavior and injuries at each workplace. The results revealed significant associations between different safety climates, safety behavior, and unintentional injuries, and provided evidence that safety behavior strongly mediates the relationship between safety climate and unintentional injuries. Our study reinforces the empirical association of occupational safety climate and safety behavior with occupational injuries and identifies some effective measures to prevent and control injuries in Chinese workplaces. Ó 2015 Elsevier Ltd. All rights reserved.

Keywords: Occupational safety climate Safety behavior Unintentional occupational injury

1. Introduction A safety climate is the sum of employees’ shared perceptions of the policies, procedures, and practices relating to safety in their work environment (Zohar, 1980; Huang et al., 2006). Although the constructs used to assess safety climate have varied from study to study, measured domains generally include management commitment, supervisor support, safety awareness, safety training, safety policy, safety knowledge, safety communication, and co-worker support (Seo et al., 2004; Olsen, 2010; O Connor et al., 2011; Brondino et al., 2012; Huang et al., 2012). Previous studies have used the theory of task and contextual performance to divide safety behavior into two types: safety participation and safety compliance (Griffin and Neal, 2000). Neal et al. (2000) firstly defined safety compliance as ‘‘complying with safety procedures and carrying out work in a safe manner,’’ and safety participation as a ‘‘safety-oriented behavior that involves the individual providing safety suggestions within the ⇑ Corresponding author. Tel.: +86 20 87332199; fax: +86 20 87330446. E-mail addresses: (W. Chen).

[email protected]

(X.

Liu),

http://dx.doi.org/10.1016/j.ssci.2015.04.023 0925-7535/Ó 2015 Elsevier Ltd. All rights reserved.

[email protected]

organization, promoting the safety program within the workplace, demonstrating initiative, and putting effort into improving safety in the workplace’’. In recent decades, it has been well documented that safety climate is related to safety behavior and unintentional injuries in workplaces in Western countries. A recent meta-analytic review revealed that safety climates offer robust predictions of objective safety criteria (the occurrence of occupational injury) and subjective safety criteria (better self-reported safety behavior) across industries (Clarke, 2006a) and countries (Christian et al., 2009). Moreover, it was found that a positive safety climate can encourage safe performance either through rewards or through principles of social exchange (Zohar, 2000; Clarke, 2006b) and that safety climate might indirectly affect safety behavior through some mediation variables (Griffin and Neal, 2000; Zohar and Luria, 2003). A number of studies showed that safety climate was directly associated with safety performance (Zohar, 2000; Siu et al., 2004; Smith et al., 2006; Clarke, 2006a; Wu et al., 2008; Brondino et al., 2012; Zohar et al., 2014). In contrast, Clarke found that safety climate was significantly related to safety behavior (i.e., safety participation and compliance), but weakly related to occupational injuries (Clarke, 2006a).

174

X. Liu et al. / Safety Science 78 (2015) 173–178

Due to China’s rapid industrialization, occupational injuries have become an important public health problem in recent decades. The Chinese Ministry of Human Resources and Social Security (2012) reported that there were approximately 1.17 million workplace injuries in 2012. However, the reason for the high occurrence of occupational injuries in China is still unclear. The present study examined whether the association between safety climate, safety behavior, and occupational injuries found in Western countries also exists in Chinese manufacturing enterprises. 2. Methods 2.1. Participants The convenience sampling method was used to select 50 manufacturing enterprises in Zhongshan City in Southern China for the study. The surveys were distributed between May 2011 and December 2012, and eight enterprises did not respond. Among the 42 enrolled enterprises, 19 were medium-size enterprise and 23 were small-size enterprises, involved in lighting processing, metal, shoes, electronics, and toys. A retrospective survey was administered to the front-line production workers of the 42 enrolled enterprises. Of the 3970 workers who received the survey, 3375 completed the whole questionnaire, giving a response rate of 85.0%. The average age of the surveyed workers was 32.67 years (SD = 8.31) with a range from 18 to 59. The average years of experience in their current job was 6.71 years (SD = 5.82). Other demographic traits are shown in Table 1.

survey, the content of the questionnaire, and how to complete it. During the data collection, the investigators answered any queries raised by the workers. All of the questionnaires were immediately checked for missing data or errors to ensure they were correctly completed.

2.3. Relevant measures 2.3.1. Occupational safety climate Building on previous occupational safety climate scales (Varonen and Mattila, 2000; Mearns et al., 2003; Cooper and Phillips, 2004; Huang et al., 2006; Lin et al., 2008; Lu and Tsai, 2008), we developed an occupational safety climate scale in Chinese. Barbaranelli’s findings provided initial empirical evidence of the cross-country measurement equivalence of the safety climate scales (Barbaranelli et al., 2015). It consisted of twenty-two items, and each item was rated on a 5-point Likert scale (1 = strongly agree, 2 = agree, 3 = neither disagree nor agree, 4 = disagree, 5 = strongly disagree). Exploratory factor analysis (EFA) was used to evaluate the safety climate scale. Variables with factor loadings of 0.45 or greater were grouped and judgments were made about their applicability to an underlying concept (Pett et al., 2003). The Kaiser–Meyer–Olkin value of 0.94 indicated that the data were suitable for factor analysis, and the Bartlett Test of Sphericity (v2 = 42766.47, p < 0.01) suggested that correlations existed between some of the response categories. A reliability test was performed to evaluate the internal consistency; the Cronbach’s alpha coefficient of 0.70 is generally accepted as the minimum desired value (Litwin, 1995). The results of the exploratory factor analysis (EFA) of the safety climate and reliability test are presented in Table 2. The factor

2.2. Data collection The selected workers were asked to complete a self-administered structured questionnaire in Chinese asking about their socio-demographic characteristics, perceived safety climate, occupational safety behavior, and work-related injuries in the past 12 months. Before the survey, several well-trained investigators explained to all of the workers the objectives of

Table 1 Demographic traits of participants. Demographic traits

n

%

Gender Male Female

2352 1023

69.7 30.3

Marital status Unmarried Married Other

835 2424 116

24.8 71.8 3.4

Educational level Primary or below Lower secondary Upper secondary Tertiary or above

366 2100 786 123

10.8 62.2 23.3 3.7

Years of work experience 65 y 6–15 y 16–25 y P26 y

1821 1324 195 35

54.0 39.2 5.8 1.0

Monthly income (RMB) 61000 1001–2000 2001–3000 3001–4000 >4000

60 984 1729 405 197

1.8 29.2 51.2 12.0 5.8

Table 2 Factor loadings of the occupational safety climate scale. Item description (variance proportion; Cronbach’s alpha coefficient)

Factor loading

Factor 1: Co-worker’s support(18.41%; 0.88) Coworkers mention safety compliance Coworkers maintain safe conditions Coworkers focus on their own work safety Coworkers comply with safety procedures Coworkers pay close attention to the safety of our team

0.802 0.797 0.716 0.685 0.662

Factor 2: Management commitment(18.12%; 0.90) Management invites employees to safety improvement sessions Management encourages employees to participate in security goal setting Management encourages safety participation Management accepts advice about safety Management is concerned about our well-being Management offers enough safety information Factor 3: Safety supervision (15.63%; 0.86) Supervisors frequently check the production situation Supervisors warn employees of no-smoking rules in the workplace Supervisors frequently check the PPE Supervisors frequently talk about safety Supervisors draw employees attention to production safety guides or warning labels Supervisors remind employees to wear PPE Safety training (10.52%; 0.73) I have been offered enough training in PPE I have been offered enough safety training during new staff induction I had an occupational health examination during new staff induction I have been offered enough training in health information I have been offered regular and useful safety training PPE: personal protective equipment.

0.766 0.760 0.753 0.658 0.632 0.492 0.713 0.701 0.681 0.628 0.492 0.473 0.697 0.557 0.556 0.506 0.469

175

X. Liu et al. / Safety Science 78 (2015) 173–178

analysis identified four factors that explained 62.7% of the total variance. According to the nature and content of the items included in each factor, the four factors were as follows: coworker support, management commitment, safety supervision, and safety training. The Cronbach’s alpha coefficients were high, ranging from 0.73 to 0.90.

2.3.2. Occupational safety behavior We used onsite observations and discussions with occupational physicians and safety managers to develop a Chinese scale to measure workers’ occupational safety behavior. Each of the twenty-one items asked a participant to state how often he followed the relevant behavior; for example, ‘‘Use the machine shield correctly.’’ Each item was rated on a 5-point Likert scale (1 = always, 2 = mostly, 3 = occasionally, 4 = never, and 5 = never encounter). The factor analysis identified three factors that explained 56.66% of the total variance. Based on the nature and contents of the items included in each factor, the three factors were defined as follows: ‘‘personal protective equipment,’’ ‘‘safety compliance,’’ and ‘‘safety initiatives.’’The Kaiser–Meyer–Olkin value was 0.90 and the Bartlett Test of Sphericity was v2 = 34630.14 (p < 0.01). The Cronbach’s alpha coefficients were high, ranging from 0.77 to 0.89 (Table 3).

2.3.3. The work-related injury experience In this study, a work-related injury was defined as any injury that occurred at the workplace; the workplace was defined as any location in which the worker was present in the course of his or her duties. All of the workers were asked whether they had been injured at work during the past 12 months. The responses were coded as 1 for ‘‘yes’’ and 0 for ‘‘No.’’ The frequency and severity of the injuries were also recorded. Injuries were categorized as serious injuries (required hospital treatment), moderate injuries (involving at least a day’s work lost), and light injuries (needed immediate treatment but did not influence work, such as minor

Table 3 Factor loadings of safety behavior scale. Item description (variance proportion; Cronbach’s alpha coefficient) Personal protective equipment (22.23%; 0.89) Wear respirators or breathing mask Wear ear defenders in noisy workplace Wear gloves as safety procedure Wear gloves while using chemicals Use the machine shield correctly Wear respirator although there is a local ventilation system in workplace Replace my respirator regularly Wear ear defenders in noisy workplace, even when talking with coworkers Safety compliance (17.23%; 0.89) Turn off the power and inform the supervisory when a machine is working abnormally Adjust posture at work to avoid fatigue Do not smoke in workplace Use a tool to adjust a dangerous part of a machine Use a cart or other hand barrow when carrying heavy goods Stop working when not offered protective equipment Read instruction manual carefully before I use a new machine Clean work bench before going off-duty Safety initiatives (17.21%; 0.77) Read material safety data sheet before using a new chemical Check the containers to prevent chemicals from leaking Cover the containers after using chemicals Inform the supervisory of chemicals without labels Use local ventilation equipment to discharge poisonous waste

bruising). In this study, only the serious and moderate injuries were included in the final data analysis. The injury frequency was transformed by logit transformation (Fullarton and Stokes, 2007) to make the data suitable for regression analyses such as path analysis. The frequency of injuries (i) was divided by the number of employees (N) for each site (g). This provided a measure of the injury risk at each site. The injury data display a Poisson distribution, and as such, did not fit a normal distribution (Hosmer and Lemeshow, 2000). The transformed log scores were then made negative to return the direction of scores, where the scores become larger when transformed. This variable, thereafter referred to as injury probability, was calculated as follows:

0

ig Ng

1

b g ¼  log @  A Y i 1  Ngg

2.4. Statistical analyses The database was constructed using EpiData 3.0. Missing data accounted for no more than 2% of any variable and were imputed using the mean replacement method. Descriptive statistics and correlations between variables were computed using SPSS 17.0 for Windows. Path analysis was performed with Amos 20.0. To examine the significance of each path, the significance levels of the regression parameters for the relationship between the variables in the estimated models (path coefficients) were denoted by the critical ratio or t-value (t-value > 1.96, p < .05), and the chi-square test was used to evaluate the fit. The model fit was considered to be good if the goodness-of-fit (GFI), the Comparative Fit Index (CFI), and the Tucker-Lewis index (TLI) were greater than 0.95, the root mean square error of approximation (RMSEA) was less than 0.05, and the normed chi-square index (v2/df ratio) was less than 3.0 (Kline, 2005). 3. Results

Factor loading

3.1. Inter-correlations between occupational safety climate, behavior, and injury

0.842 0.818 0.763 0.763 0.728 0.710

The descriptive statistics and inter-correlations between the latent variables of occupational safety climate, behavior, and work-related injury are presented in Table 4. Except for the correlations between ‘‘coworker support’’ and ‘‘personal protective equipment,’’ between ‘‘safety supervision’’ and ‘‘personal protective equipment,’’ and between ‘‘safety training’’ and ‘‘work-related injuries,’’ the correlations between the variables were significant with correlation coefficients ranging from 0.068 to 0.712.

0.706 0.686

0.736 0.705 0.592 0.588 0.588 0.584 0.548 0.544 0.839 0.824 0.797 0.697 0.587

3.2. Path analysis The path model fit the data well (GFI = 0.999; CFI = 0.996; TLI = 0.992; RMSEA = 0.011; v2/df = 1.385). The path coefficients in the path diagrams showed that the occupational safety climates may affect the work-related injuries indirectly through occupational safety behavior (Fig. 1). Both ‘‘safety supervision’’ and ‘‘management commitment to safety’’ reduced the work-related injuries by improving ‘‘personal protective equipment use,’’ ‘‘safety compliance,’’ and ‘‘safety initiatives.’’‘‘Coworkers support’’ decreased the work-related injury by promoting ‘‘personal equipment use’’ and ‘‘safety compliance,’’ whereas ‘‘safety training’’ lessened the

176

X. Liu et al. / Safety Science 78 (2015) 173–178

Table 4 Inter-correlation matrix for safety climate, safety behavior, and work-related injury.

CS MC SS ST PPE SI SC INJ

Mean ± SD

CS

MC

SS

ST

PPE

SI

SC

1.998 ± 0.468 2.099 ± 0.570 1.871 ± 0.484 2.216 ± 0.517 1.376 ± 0.768 1.833 ± 0.818 2.258 ± 1.375 N/A

0.695** 0.619** 0.485** 0.022 0.169** 0.213** 0.068**

0.712** 0.540** 0.068** 0.232** 0.268** 0.155**

0.540** 0.016 0.324** 0.293** 0.201**

0.099** 0.174** 0.157** 0.020

0.192** 0.298** 0.160**

0.629** 0.243**

0.267**

N/A: not applicable. MC: management commitment; SS: safety supervision; CS: coworker support; ST: safety training; PPE: personal protective equipment; SI: safety initiatives; SC: safety compliance; INJ: injury. ** p < 0.01.

SS PPE

CS

SC INJ

MC

SI

ST Fig. 1. Path model with occupational safety climates as predictors, occupational safety behavior as mediators, and work-related injuries as outcomes. MC: management commitment; SS: safety supervision; CS: coworker support; ST: safety training; PPE: personal protective equipment; SI: safety initiatives; SC: safety compliance; INJ: injury ⁄⁄⁄ p < 0.01. ⁄p < 0.05.

Table 5 Direct, indirect, and total effects of occupational safety climate and behavior on work-related injury. ST SI Direct effect Indirect effect Total effect SC Direct effect Indirect effect Total effect PPE Direct effect Indirect effect Total effect Work-related injury Direct effect Indirect effect Total effect

MC

CS

SS

0.114**

0.301**

0.114**

0.301**

0.169**

0.119**

0.127**

0.169**

0.119**

0.127**

0.121**

0.064**

0.043*

0.116**

0.121**

0.064**

0.043*

0.116**

0.018 0.018

0.049 0.049

0.011 0.011

0.041 0.041

SI

SC

PPE

0.130**

0.145**

0.146**

0.130**

0.145**

0.146**

MC: management commitment; SS: safety supervision; CS: coworker support; ST: safety training; PPE: personal protective equipment; SI: safety initiatives; SC: safety compliance. * P < 0.05. ** P < 0.01.

work-related injuries by increasing the use of ‘‘personal protective equipment.’’ Table 5 shows the estimates of the standardized direct, indirect, and total effects of occupational safety climates and behavior on

work-related injuries. Of all the standardized total effects, the absolute value of the use of ‘‘personal protective equipment’’ was the greatest, followed by ‘‘safety compliance,’’ ‘‘safety initiative,’’ ‘‘management commitment,’’ ‘‘safety supervision,’’ ‘‘safety

X. Liu et al. / Safety Science 78 (2015) 173–178

training,’’ and ‘‘coworker support.’’For the standardized direct effects, the absolute value of the use of ‘‘personal protective equipment’’ was the greatest, followed by ‘‘safety compliance’’ and ‘‘safety initiative.’’ For the standardized indirect effects, the absolute value of the use of ‘‘management commitment’’ was the greatest, followed by ‘‘safety supervision,’’ ‘‘safety training,’’ and ‘‘coworker support.’’

4. Discussion The results of this study suggest that safety climate predicts safety behavior, and that safety behavior mediates the relationship between safety climate and occupational injury in China. These findings are in line with previous studies in other countries (Hofmann and Stetzer, 1996; Cavazza and Serpe, 2009). In their meta-analysis, Christian et al. (2009) developed a theoretical model in which safety climate was the antecedent of safety performance and had an indirect effect on occupational injury through safety performance. The study of mediation factors is important because it allows us to understand the mechanisms through which safety climate operates on workers’ behavior and reduces the risk of injury. Our study established empirical links between specific dimensions of safety climate, such as safety supervision (SS), management commitment (MC), coworker support (CS), safety training (ST), and safety behavior, such as safety initiates (SI), personal protective equipment (PPE), and safety compliance (SC). SS and MC had a positive relationship with three different dimensions of safety behavior. Safety initiatives (SI), personal protective equipment (PPE), and safety compliance (SC) had similar associations with occupational injury. Our results revealed that safety supervision (SS) was the most proximal antecedent of safety initiative behavior, which also increased the use of PPE and employees’ safety compliance. Zohar and Luria (2003) carried out an intervention in which supervisors were trained to participate in safety-oriented communication with subordinate workers; they found that the intervention reduced minor injury rates and increased PPE use. In a study of the maintenance of heavy duty equipment, Zohar (2002) found that improved communication between line workers and supervisors resulted in decreasing micro accidents and increasing PPE use. Sampson suggested that supervisors’ use of communication had a positive effect on safety compliance and safety participation (Sampson et al., 2014). According to the Leader–Member–exchan ge (LMX) theory (Graen and Uhl-Bien, 1995), supervisors can increase safety initiatives among employees by inspiring safety and improving PPE use, and they can increase employees’ safety compliance by using safety monitoring (Griffin and Hu, 2013). Our results expanded upon these findings by suggesting that communication between supervisors and workers might also improve the safety behavior of workers. Coworker support (CS) was also an important dimension of safety climate. Our results showed that coworker support was positively associated with SC and PPE (b = 0.119, p < 0.001; b = 0.043, p < 0.05). Work group safety mainly involved coworkers’ involvement in and commitment to safety issues: if an employee perceived that his/her coworkers were concerned about safety, the whole group tended to practice safe behavior (Hayes et al., 1998). In a study of manufacturing industries in Italy, researchers showed that coworkers’ safety climate had a stronger influence on safety behavior, especially on safety participation, than supervisors’ safety climate (Brondino et al., 2012). However, in our study, the path CS–SI was non-significant, indicating that in Chinese manufacturing groups, the mutual influence between colleagues is not sufficient to encourage workers to participate in active safety behavior. It is necessary for higher-level managers, using

177

techniques such as increasing support with incentive policies to strengthen supervision and promote safe behavior. Management commitment (MC) (b = 0.169, p < 0.001) was the most proximal antecedent of safety compliance behavior and had an indirect effect on occupational injury. Like the SS dimension, MC also had a significant effect on three dimensions of safety behavior. These results were consistent with previous scholars’ conclusions that MC is a major factor in safety climate and is critical to employee safety performance (Christian et al., 2009). Kao used a structural equation model to test the relationship between management commitment and safety performance, including safety initiatives, compliance behavior, and occupational injury (Kao et al., 2009). Their findings revealed that MC had no direct correlation with compliance behavior and injury, but a positive correlation with SI, which was partly consistent with the results of our study. MC assesses the extent to which people perceive that management values safety, how they communicate safety issues, and their actions that support safety. MC is a sign of the extent to which top management demonstrates positive and supportive attitudes toward worker safety. In China, when top management is committed to safety, they may set goals for safe production, establish an occupational safety management organization, increase communication and feedback, and provide enough support and resources for safety activities. All of these measures help increase the three types of safety behavior. Safety training was positively associated with the use of personal protective equipment (b = 0.121, p < 0.001). However, in the present study the ‘‘SS–SI’’ and ‘‘SS–SC’’ paths were non-significant, indicating that safety training did not significantly alter safety behavior encourage employees to comply with safety procedures. This is not consistent with previous studies (Cooper and Phillips, 2004; Huang et al., 2012), which suggested that safety training (one dimension of the safety climate) could predict the actual levels of safety behavior among manufacturing employees. Our results may reflect the fact that in China, safety training is mainly for new employees and aimed at improving their safety knowledge (i.e., operating the machine correctly and wearing PPE). Other types of daily safety training are conducted sporadically and have limited effects on safety behavior. Personal protective equipment, safety compliance, and initiative behavior were all associated with lower rates of self-reported occupational injury (b = 0.146, b = 0.145, b = 0.130, p < 0.001). Our findings about EFA support the trend in the occupational safety research literature toward conceptualizing safety behavior as multi-dimensional (Griffin and Neal, 2000; MartínezCórcoles et al., 2011; Zhang and Wu, 2014). For example, in this study, personal protective equipment was distinguished from safety compliance behavior as in the other recent studies (Cavazza and Serpe, 2009; Arcury et al., 2012; Tholén et al., 2013). According to the American National Safety Council, failure to use protective gear provided at the workplace accounts for approximately 40% of work accidents (Olson et al., 2009). Safety compliance refers to voluntary behavior that is required to carry out normal operations and safety initiatives. PPE, SC, and SI had similar effects on lowering the occurrence of occupational injury, suggesting that future interventions in manufacturing industries should pay more attention to the three dimensions of safety behavior. Although this study extends the existing safety climate literature by demonstrating several safety climate–behavior–injury relationships, it has some limitations worth mentioning. First, like other safety climate studies, the results of this study were derived from a cross-sectional survey, preventing us from making definitive causal conclusions due to the nature of the cross-sectional design. In further field studies, a longitudinal design could be used to clarify these associations. We used a self-reported questionnaire

178

X. Liu et al. / Safety Science 78 (2015) 173–178

to measure occupational safety climate, behavior, and injury, although the injury reports were confirmed using the companies’ medical records, the data need to be interpreted cautiously. In addition, as a result of the limitations in the data collection, focus was placed on only four main dimensions of safety climate in the manufacturing industries; other dimensions such as safety motivation and risk perception were not considered and should be further investigated (Neal and Griffin, 2006; Ma and Yuan, 2009). Furthermore, as a chemical plant is different from a laundry, a logistics company, or manufacturing company, intercompany analysis should be investigated in future studies. In summary, occupational safety climate and safety behavior in the workplace are related to the occurrence of occupational injury in Chinese manufacturing factories. Improving the occupational safety climate and safety behavior may decrease the number of occupational injuries in the workplace. Acknowledgements This study was supported with funds from the following sources: The China Medical Board (Grant No. 08-924), Guangdong Provincial Science and Technology Department Projects Fund (Grant No. A2013B021800034), and Zhongshan Health Major Science and Technology Project (Grant No. 20132A003). The authors express their deep gratitude to Benjamin Anderson for the help of the English Language Editing and the Zhongshan CDC, which helped with data collection. References Arcury, T.A., Grzywacz, J.G., Anderson, A.M., Mora, D.C., Carrillo, L., Chen, H., Quandt, S.A., 2012. Personal protective equipment and work safety climate among Latino poultry processing workers in Western North Carolina, USA. Int. J. Occup. Environ. Health 18 (4), 320–328. 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. 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. 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. Saf. Res. 40 (4), 277–283. 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. 94 (5), 1103–1127. Clarke, S., 2006a. The relationship between safety climate and safety performance: a meta-analytic review. J. Occup. Health Psychol. 11 (4), 315–327. Clarke, S., 2006b. Contrasting perceptual, attitudinal and dispositional_x000a_approaches to accident involvement in the workplace. Saf. Sci. 44, 537–550. Cooper, M.D., Phillips, R.A., 2004. Exploratory analysis of the safety climate and safety behavior relationship. J. Saf. Res. 35 (5), 497–512. Fullarton, C., Stokes, M., 2007. The utility of a workplace injury instrument in prediction of workplace injury. Accid. Anal. Prev. 39 (1), 28–37. Graen, G.B., Uhl-Bien, M., 1995. Relationship-based approach to leadership: Development of leader-member exchange (LMX) theory of leadership over 25 years: applying a multi-level multi-domain perspective. Leadership Q. 6 (2), 219–247. Griffin, M.A., Hu, X., 2013. How leaders differentially motivate safety compliance and safety participation: the role of monitoring, inspiring, and learning. Saf. Sci. 60, 196–202. 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. Hayes, B.E., Perander, J., Smecko, T., Trask, J., 1998. Measuring perceptions of workplace safety: development and validation of the work safety scale. Accid. Anal. Prev. 29 (3), 145–161. Hofmann, D.A., Stetzer, A., 1996. A cross-level investigation of factors influencing unsafe behaviors and accidents. Personnel Psychol. 49 (2), 307–339.

Hosmer, D.W., Lemeshow, S., 2000. Applied Logistic Regression, second ed. John_x000a_Wiley and Sons, Canada. Huang, Y., Ho, M., Smith, G.S., Chen, P.Y., 2006. Safety climate and self-reported injury: assessing the mediating role of employee safety control. Accid. Anal. Prev. 38 (3), 425–433. Huang, Y., Verma, S.K., Chang, W., Courtney, T.K., Lombardi, D.A., Brennan, M.J., Perry, M.J., 2012. Management commitment to safety vs. employee perceived safety training and association with future injury. Accid. Anal. Prev. 47, 94–101. Kao, L., Stewart, M., Lee, K., 2009. Using structural equation modeling to predict cabin safety outcomes among Taiwanese airlines. Transport. Res. Part E: Logis. Transport. Rev. 45 (2), 357–365. Kline, R.B., 2005. Principles and Practice of Structural Equation Modeling. TheGuilford Press, New York. Lin, S., Tang, W., Miao, J., Wang, Z., Wang, P., 2008. Safety climate measurement at workplace in China: a validity and reliability assessment. Saf. Sci. 46 (7), 1037– 1046. Litwin, M.S., 1995. How to Measure Survey Reliability and Validity. Sage, Thousand_x000a_Oaks. Lu, C., Tsai, C., 2008. The effects of safety climate on vessel accidents in the container shipping context. Accid. Anal. Prev. 40 (2), 594–601. Ma, Q., Yuan, J., 2009. Exploratory study on safety climate in Chinese manufacturing enterprises. Saf. Sci. 47 (7), 1043–1046. Martínez-Córcoles, M., Gracia, F., Tomás, I., Peiró, J.M., 2011. Leadership and employees’ perceived safety behaviours in a nuclear power plant: a structural equation model. Saf. Sci. 49 (8–9), 1118–1129. Mearns, K., Whitaker, S.M., Flin, R., 2003. Safety climate, safety management practice and safety performance in offshore environments. Saf. Sci. 41 (8), 641– 680. Neal, A., Griffin, M.A., 2006. A study of the lagged relationships 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. O Connor, P., O Dea, A., Kennedy, Q., Buttrey, S.E., 2011. Measuring safety climate in aviation: a review and recommendations for the future. Saf. Sci. 49 (2), 128– 138. Olsen, E., 2010. Exploring the possibility of a common structural model measuring associations between safety climate factors and safety behaviour in health care and the petroleum sectors. Accid. Anal. Prev. 42 (5), 1507–1516. Olson, R., Grosshuesch, A., Schmidt, S., Gray, M., Wipfli, B., 2009. Observational learning and workplace safety: the effects of viewing the collective behavior of multiple social models on the use of personal protective equipment. J. Saf. Res. 40 (5), 383–387. Pett, M.A., Lackey, N.R., Sullivan, J.J., 2003. Making sense of factor analysis: the use of factor analysis for instrument development in health care research. Sage. Sampson, J.M., DeArmond, S., Chen, P.Y., 2014. Role of safety stressors and social support on safety performance. Saf. Sci. 64, 137–145. Seo, D.C., Torabi, M.R., Blair, E.H., Ellis, N.T., 2004. A cross-validation of safety climate scale using confirmatory factor analytic approach. J. Saf. Res. 35 (4), 427–445. Siu, O.L., Phillips, D.R., Leung, T.W., 2004. Safety climate and safety performance among construction workers in Hong Kong. The role of psychological strains as mediators. Accid. Anal. Prev. 36 (3), 359–366. Smith, G.S., Huang, Y.H., Ho, M., Chen, P.Y., 2006. The relationship between safety climate and injury rates across industries: the need to adjust for injury hazards. Accid. Anal. Prev. 38 (3), 556–562. Tholén, S.L., Pousette, A., Törner, M., 2013. Causal relations between psychosocial conditions, safety climate and safety behaviour – a multi-level investigation. Saf. Sci. 55, 62–69. Varonen, U., Mattila, M., 2000. The safety climate and its relationship to safety practices, safety of the work environment and occupational accidents in eight wood-processing companies. Accid. Anal. Prev. 32 (6), 761–769. Wu, T., Chen, C., Li, C., 2008. A correlation among safety leadership, safety climate and safety performance. J. Loss Prevent. Proc. 21 (3), 307–318. Zhang, J., Wu, C., 2014. The influence of dispositional mindfulness on safety behaviors: a dual process perspective. Accid. Anal. Prev. 70, 24–32. 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 microaccidents in manufacturing jobs. J. Appl. Psychol. 85 (4), 587– 596. Zohar, D., 2002. Modifying supervisory practices to improve subunit safety: a leadership-based intervention model. J. Appl. Psychol. 87 (1), 156–163. Zohar, D., Luria, G., 2003. The use of supervisory practices as leverage to improve safety behavior: a cross-level intervention model. J. Saf. Res. 34 (5), 567–577. Zohar, D., Huang, Y., Lee, J., Robertson, M., 2014. A mediation model linking dispatcher leadership and work ownership with safety climate as predictors of truck driver safety performance. Accid. Anal. Prev. 62, 17–25.