Supervisory interpretation of safety climate versus employee safety climate perception: Association with safety behavior and outcomes for lone workers

Yueng-Hsiang Huang; Michelle M. Robertson; Jin Lee; Jenn Rineer; Lauren A. Murphy; Angela Garabet; Marvin J. Dainoff

Contents lists available at ScienceDirect

Transportation Research Part F

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

Supervisory interpretation of safety climate versus employee safety climate perception: Association with safety behavior and outcomes for lone workers

Yueng-hsiang Huang a'*, Michelle M. Robertson a, Jin Leea,b, Jenn Rineera,c, Lauren A. Murphy a,d, Angela Garabeta, Marvin J. Dainoffa

a Liberty Mutual Research Institute for Safety, Hopkinton, MA, USA b University of Connecticut, Storrs, CT, USA c Portland State University, Portland, OR, USA d Harvard School of Public Health, Boston, MA, USA

ARTICLE INFO ABSTRACT

Research has shown that safety climate predicts safety behavior and safety outcomes in a variety of settings. Prior studies have focused on traditional work environments in which employees and supervisors work in the same location and the mechanisms through which safety climate affects behavior are largely understood. However, the nascent research examining safety climate among lone workers suggests that safety climate may have some uniqueness in this context. Based on leadership theories and utilizing an exploratory approach, this study increases our understanding of the lone worker context by examining employee perception of safety climate and supervisory interpretation of safety climate; how similar or different they are, and how they are related to important safety outcomes. Surveys were administered to a matched sample of 1831 truck drivers and their 219 supervisors at four different trucking companies. Objective data on employee injuries were collected six months after survey administration. The results provided support for the measurement equivalence of the Trucking Safety Climate Scale at the organization level for both employee and supervisor respondents. For both organization- and group-level safety climate, employee perceptions of safety climate and supervisory interpretation of safety climate were significantly different, such that supervisors provided higher ratings for both safety climate sub-scales. Further, only employee safety climate perceptions significantly predicted self-reported safety behavior (directly) and objective injury outcomes (indirectly). This suggests that when trying to gauge and improve upon a trucking company's safety climate, we should rely on employee perspectives, rather than supervisory interpretation, of safety climate.

BY license (http://creativecommons.org/licenses/by/3XI/).

CrossMark

Article history:

Received 25 March 2013

Received in revised form 28 March 2014

Accepted 2 April 2014

Available online 3 July 2014

Keywords: Safety climate Trucking industry

Supervisor versus employee perceptions Workplace accidents and injuries

1. Introduction

Occupational safety is an issue with significant financial and societal consequences. In 2011, there were over 4600 fatal workplace injuries in the U.S. (Bureau of Labor Statistics, 2012). According to the 2012 Liberty Mutual Workplace Safety

* Corresponding author. Address: Center for Behavioral Sciences, 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).

This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/).

Index, the most disabling workplace injuries and illnesses in 2010 amounted to $51.1 billion in direct U.S. workers' compensation costs (Liberty Mutual Research Institute for Safety, 2012). When considering both direct and indirect costs of all occupational injuries and illnesses, estimates are as high as $250 billion annually (Leigh, 2011).

While there has been a resurgence in workplace safety research in recent years, there are still many gaps that need to be addressed in order to improve the health and safety of workers. One major gap regards how to improve safety among lone workers (e.g., truck drivers), defined as employees who work alone and perform an activity that is intended to be carried out in isolation from other workers, without close or direct supervision (British Security Industry Association, 2010; Hughes & Ferrett, 2009). There is reason to believe that the mechanisms through which safety climate influences behavior are different among lone workers as compared to those in a traditional work environment. In developing and validating safety climate scales for lone workers (using truck drivers as an exemplar), Huang et al. (2013) found that truck drivers' safety climate perceptions predicted drivers' safety behaviors and future road injuries, and their industry-specific items provided stronger predictive value in safe driving behavior than did generic safety climate items. This suggests that taking the context of the work environment into account is important in understanding safety climate for lone workers (specifically, truck drivers).

In the Huang et al. study (2013), the authors found that, contrary to the established conceptualization of safety climate (e.g., Zohar, 1980, 2000, 2011), there were no shared safety climate perceptions among the groups of lone workers that they studied. This implies that in the unique context of trucking, something other than shared perceptions with peers may be informing employees' safety climate perceptions. A study from Zohar, Huang, Lee, and Robertson (2014) showed that dispatcher (distant) leadership (measured by the Leader-Member Exchange scale; Graen & Uhl-Bien, 1995) is a significant antecedent of truck drivers' safety climate perceptions, driving behaviors, and hard-braking frequencies (i.e., accident near-miss events). Thus, absent frequent contact with co-workers, the dispatcher (supervisor), despite her/her remoteness, is likely to be the primary conduit through which the employee picks up the informational elements upon which to base his/ her safety climate perceptions. In turn, these informational elements are presumably based on the dispatcher's own interpretation of organizational-level safety climate and his/her own values and attitudes. For in-house workers, the supervisor's role in this transmission would be supplemented by co-workers.

Continuing this line of lone worker/truck driver research, and based on leadership theories, the current study compares supervisory interpretation of safety climate and employee safety climate perceptions in order to better understand the antecedents of lone worker safety, which ultimately impacts accident and injury outcomes. Specifically, this study makes the following three contributions: (1) it demonstrates measurement equivalence (ME) of the organization-level safety climate sub-scale items among employee and supervisor respondents. Establishment of ME allows meaningful and fair comparisons of ratings on a single scale from multiple parties; (2) it is the first study to explore the similarity or difference between supervisory interpretation of safety climate versus employee safety climate perception for lone workers (using dispatcher and truck drivers as exemplar, respectively), for both organizational-level and group-level safety climate perceptions; and (3) it expands prior research by testing simultaneously the impact of supervisory interpretation of safety climate and employee safety climate perception on safety outcomes.

1.1. Safety climate

Research has shown that safety climate (traditionally defined as workers' shared perceptions of their organization's policies, procedures, and practices as it relates to the value and importance of safety within the organization; Griffin & Neal, 2000; Zohar, 1980, 2000, 2011) predicts safety behavior and safety outcomes (such as accidents and injuries) in a variety of settings (e.g., Beus, Payne, Bergman, & Arthur, 2010; Christian, Bradley, Wallace, & Burke, 2009). It is a multilevel construct comprising two levels: organization-level safety climate (employees' perceptions of the company's commitment to and prioritization of safety) and group-level safety climate (employees' perceptions of their direct supervisors' commitment to and prioritization of safety) (e.g., Zohar, 2008). While similar to safety culture, safety climate is a distinct construct. Safety culture is defined as shared values and beliefs that interact with an organization's structure and control systems to produce behavioral norms (Reason, 1998; Thompson, 1996). Safety climate, on the other hand, focuses on workers' perceptions. In other words, safety climate can be viewed as a measurable marker of safety culture (e.g., Huang et al., 2013).

1.2. The need for studying safety climate among lone workers/truck drivers

A ''lone worker'' can be anyone who works on his/her own, either regularly or occasionally, without access to immediate support from coworkers or managers (National Health Service (NHS), 2005). In today's workforce, an increasing number of employees perform work assignments away from the traditional office setting, either because the type of work they perform requires them to do so, or because advances in technology have allowed for flexibility in work location (Golden, Veiga, & Dino, 2008). While professional isolation has only recently come into the spotlight in the behavioral science literature, some research has begun to suggest that remote workers may feel ''out of the loop'' (Baruch & Nicholson, 1997) and experience a decreased sense of belonging (Baumeister & Leary, 1995), among other negative outcomes. However, other studies have shown that there can be a positive side to working remotely. For example, Gajendran and Harrison (2007) found that isolated workers experienced increased perceived autonomy and lower work-family conflict. They also found that telecommuting did not have detrimental effects on workplace relationships (except for those who telecommuted for the majority of their workweek).

While there has been a dramatic increase in the number of safety climate studies published in recent years (e.g., Beus et al., 2010; Christian et al., 2009; Huang, Chen, & Grosch, 2010; Zohar, 2010), most have focused on traditional work environments. Traditional employees, also referred to as in-house workers, have many opportunities throughout the day to interact with one another and synthesize their first-hand experiences of formal and informal policies and procedures; however, as mentioned above, some employees work away from a traditional office setting (e.g., telecommuters or those working in the trucking, utility, and airline industries). Because lone working is becoming increasingly common, and relatively little is known about workplace safety attitudes and behaviors in this context, it is important that safety climate research begin to address lone workers. The current study focuses on lone workers and uses long-haul truck drivers as an example.

The trucking transportation industry is inherently dangerous. In 2010, there were 396 fatalities in the trucking industry (Bureau of Labor Statistics, 2012). The same data also showed that the trucking industry experienced nearly ten times more fatalities (with a rate of 31.8 per 100,000 workers) than the cross-industry average (with a rate of 3.6 per 100,000 workers). Furthermore, non-fatal injuries occur at a rate of 5.3 per 100 full-time truck drivers while the average across private industries is 2.9 per 100 workers (Bureau of Labor Statistics, 2012). Finally, accidents in the trucking industry can involve other commercial or passenger vehicles and property, compounding the negative impact of these events; this showcases the need for further study of accident prevention in this industry.

1.3. Gaps in literature related to supervisory interpretation of safety climate

Because safety climate is defined as employees' shared perceptions, prior studies have mainly focused on only front-line employee safety climate perceptions. There is interest, however, in exploring views of other members of organizations (i.e., managers and supervisors; see Cheyne, Tomas, Cox, & Oliver, 2003; Cox, Tomas, Cheyne, & Oliver, 1998; Huang et al., 2012; Zohar & Luria, 2010), in addition to the general employees, in order to better understand the antecedents of worker safety and the impact on accident and injury outcomes. It is suggested that exploring the similarities and differences of safety attitudes and climate between occupational levels is particularly important for planned improvements that target all employees at all levels (e.g., Cheyne et al., 2003). Since safety climate is traditionally defined with respect to employees' perceptions, the current study utilized the approach of Zohar and Luria (2010) by defining supervisors' responses to survey items relating to organization safety priorities and commitment to safety as supervisory interpretation of safety climate.

Prior research, however, has produced inconclusive or contradictory findings between occupational levels (i.e., management and workforce) within the same organization in terms of the interpretations of safety climate. Some studies have suggested that there may be a positive relationship between a leader's own attitudes and perception of safety and his/ her employees' safety climate perception. Drawing from the literature, there are several potential reasons for the similarity.

1.3.1. Similarity between supervisory interpretation and employee perception: organizational-level safety climate

It has often been said in the literature that ''leaders create climate'' (Lewin et al., 1939). This is consistent with the research that shows that management commitment to safety is the most important dimension of safety climate (e.g., Zohar, 1980, 2000). A key feature of safety climate is the equivalence, or parity, of espoused and enacted policies, meaning that what managers say and what they do are in alignment (Argyris & Schon, 1996; Simons, 2002). The extent to which such alignment exists at the top management/organizational level should be perceived by supervisors and, in turn, transmitted to their direct reports. Zohar and Luria (2010) found, in fact, that supervisors can affect organizational climate perceptions of group members through their role as communicators and interpreters of climate (called informing behavior or mediating behavior); therefore, supervisors' individual assessments and members' assessments of the same organizational climate cues should be correlated (Zohar & Luria, 2010). This impact should be particularly salient in the case of lone workers given the absence of frequent interaction of employees with co-workers (Zohar et al., 2014).

1.3.2. Similarity between supervisory interpretation and employee perception: group-level safety climate

Supervisors create climate for their work groups (leaders create climate). If what supervisors say and what they do are in alignment, it is very likely that supervisor and employee would have a similar interpretation of group-level safety climate. Such supervisory impact should, in turn, result in the development of shared mental models for both group members and supervisors regarding their experiences in the workplace (Rouse & Morris, 1986). Mental models, or organized knowledge structures, allow people to interact with their environment by helping them to predict and explain behavior, recognize and remember relationships among components in the environment, and construct expectations regarding what is likely to occur next (Rouse & Morris, 1986). According to Stout, Cannon-Bowers, and Salas (1996), developing shared mental models among employees is most important in situations where coworkers cannot communicate freely with one another, since they cannot directly discuss their job demands and issues. Rather, they rely on these schemas of information to make sense of their environment. To the extent that mental models of supervisors and their employees contain overlapping elements of knowledge regarding safety-related policies, practices and procedures, supervisory interpretation of safety climate and employee safety climate perception should be similar. Moreover, we should expect not only an absence of statistically significant differences between average supervisor and employee scores, but we should also expect that the capability to predict safety outcomes should be similar.

1.3.3. Differences between supervisory interpretation and employee perception: organizational-level safety climate

On the other hand, situations that lead to disharmony in interpretations of safety climate between supervisors and their employees have also been identified in the literature. Differences have been uncovered between occupational levels (i.e., management and workforce) within the same organizations in terms of the interpretations of safety attitudes and climate (e.g., Cox et al., 1998; Cheyne et al., 2003; Huang et al., 2012).

1.3.3.1. Gate-keeper model. Conceptually, it is possible to understand such differences by considering the gatekeeper function of the supervisor (Zohar & Luria, 2010). As such, the supervisor's mental model contains knowledge structures, parts of which are constructed from his/her perception of top management policies. These knowledge structures include procedural as well as declarative components (Blickensdorfer et al., 2000). The supervisor's everyday interactions with direct reports will be, to a great extent, influenced by the content of his/her mental models. In turn, the informational elements contained in these interactions will, for the case of lone workers, form the primary basis of employee perceptions of safety climate.

If the supervisor's perception of the degree of parity between espoused and enacted policies (Eakin, 1992; Paté-Cornell, 1990; Wright, 1986) of top management is discrepant with his/her own beliefs and actions, the result could be a source of employee-supervisor differences in organization-level safety climate. One way this might occur is if top management policies lacked parity (i.e., safety is important but productivity always comes first), but the gatekeeping function of the supervisor was one of mitigation; developing procedures to protect employee safety. In this case, the supervisor would (realistically) interpret the organizational safety climate to be lower than the employees' perceptions which are filtered through the more positive actions of the supervisor. In another example, top management policy might well be one of enacted-espoused parity, whereas the supervisor is willing to disregard safety policies/procedures when production falls behind. Here, the supervisor would interpret the organizational safety climate as higher than the employees' perceptions which are now filtered through the less positive actions of the supervisor. In both cases, the mechanism underlying the supervisor-employee difference is a functional discrepancy in the content of the supervisor mental model between those knowledge structures representing his/her understanding of management policy and a different set of knowledge structures representing operational day to day procedures and practices. However, the employees' mental model is primarily derived from interaction with the supervisor.

1.3.3.2. Opportunity to observe. An additional possibility for explaining supervisor-employee differences is that supervisors have greater opportunity to observe and have more frequent interaction with upper levels of management (e.g., Lawler, 1967). Cheyne et al. (2003), collecting data from two manufacturing companies, found that the different work levels exhibited different interpretations of safety climate, reflecting the view from their levels. Managers had the most positive views, followed by supervisors, and then general employees, who had the least positive views. Cheyne et al. (2003), however, did not link these perceptions to safety outcomes. If, in our own study, supervisors, being closer to the top, are better able to discern the true priorities of upper management (e.g., safety over competing demands), then supervisor scores for organization-level safety climate may be stronger predictors of safety performance than employee perceptions and more related to a company's safety outcomes.

1.3.3.3. Social desirability. Alternatively, social desirability bias (e.g., Maccoby & Maccoby, 1954) may explain the organization level differences between supervisors' and their employees' perceptions. Supervisors may identify with upper management in their companies and may alter their answers to enhance the company image. Respondents are often unwilling or unable to report accurately on sensitive topics and the result is data that are systematically biased toward respondents' perceptions of what is "correct" or socially acceptable, so-called social desirability bias (e.g., Maccoby & Maccoby, 1954).

1.3.4. Differences between supervisory interpretation and employee perception: group-level safety climate

1.3.4.1. Discrepancy between espoused and enacted policies at group level. If significant differences exist between group-level employee perceptions of supervisors and the supervisors self-assessments of safety climate, it is possible that lack of parity between espoused and enacted policies, discussed earlier with reference to top management, extends to the group level. For example, a given supervisor may be aware of, and believe he/she supports, management policy against employees' use of cell phones while driving (declarative knowledge component of supervisor mental model). At the same time, if production pressures are high, the supervisor may judge the situation to be "exceptional" and call the driver on the road (procedural knowledge component of supervisor mental model). The supervisor may not be aware of this discrepancy and, while answering the safety climate survey, interprets his/her own behavior as consistent with his/her declarative knowledge of management policy as well as aligned with messages he/she thinks he/she has sent to the work group. However, the messages received by group members, which form the primary basis for their safety climate perception, may not be what the supervisor thought he/she sent (e.g., "exceptions" occur more frequently than the supervisor realizes). Role theory (Katz & Kahn, 1978; Merton, 1957) distinguishes between sent and received (i.e., interpreted) role expectations and the fact that the two might be misaligned due to the idiosyncrasies of interpersonal communication.

1.3.4.2. Self-assessment bias. Alternatively, studies of multi-source feedback and performance ratings (e.g., employee versus supervisor sources) have repeatedly shown inconsistencies among raters indicating that self-appraisals are higher than

other-appraisals (self-assessment bias; Mount, 1984a,b; Borman, 1997). In these situations, it is possible that there is a gap between supervisory interpretation of safety climate versus employee safety climate perceptions, and supervisors may give higher scores on the safety climate survey.

1.3.5. Significance of study outcomes

The literature has shown that employees' safety climate perception is one of the best predictors of organizational safety outcomes. If evidence can be shown that leaders' interpretations of safety climate are similar their employees, it may imply that leaders' interpretation of organizational safety climate could be a good leading indicator of safety outcomes as well. If our evidence shows no similarity, this is still useful information. Prior studies (e.g., Cheyne et al., 2003) have suggested that the fact that management sees things differently from the general workforce is important in terms of promoting a positive and appropriate culture for safety. It has implications for the success of improvement programs aimed at all employment levels.

To fill a gap in the literature, the present study aims to investigate the etiology and consequences of safety climate in the trucking industry by examining supervisory interpretation of safety climate and employee safety climate perceptions; how similar or different they are, and how they are related to important safety outcomes.

1.4. Comparing supervisory interpretation of organization-level safety climate and employee organization-level safety climate perception

Since the organization-level survey portion was identical for both employees and supervisors, with the company (top management) as the referent for both, our first objective (Objective 1) was to determine whether the organization-level safety climate sub-scale had the same measurement implications among employee and supervisor respondents (in other words, to test measurement equivalence of the organization-level safety climate sub-scale). The literature has suggested that a psychological measure can be interpreted in considerably different ways across different groups of people because particular concepts used in the measure and item wordings might be understood differently across the groups. Moreover, the reference point for scale item endorsement can differ across groups. For example, on a 5-point Likert scale, a score of three can be viewed as high or good enough in certain groups, whereas the same rating can be viewed as below average or poor in other groups. Thus, measurement equivalence (Raju, Laffitte, & Byrne, 2002) should be established in order to make meaningful and fair comparisons of ratings on a single scale from multiple parties. If measurement equivalence was observed, this would imply that it is appropriate to compare supervisor and employee scores (and that the meaning of responses is the same for both groups).

If measurement equivalence can be established, Objective 2 would be to explore whether or not employee and supervisor scores on the organization-level safety climate sub-scale were statistically similar or different. If these scores are statistically different, the next step would be to determine which was more predictive of safety behavior and injury severity: supervisory interpretation of organization-level safety climate or employee organization-level safety climate perceptions (Objective 3). Though it seems likely that employee perceptions would be more strongly related to employees' own safety behavior and injury severity, the lone work context may again affect this relationship. It is possible that because supervisors are more informed of upper management's commitment to safety, they have a better understanding of organizational factors that could impact safety outcomes.

1.5. Comparing employee group-level safety climate perceptions with supervisor self-appraisals

We further explored employees' ratings versus supervisory interpretation of group-level safety climate. For the group-level safety climate sub-scale, the employees and supervisors had different questions. The referent was the supervisor in both cases, such that employees were providing an upward-appraisal of their supervisor, while supervisors were providing a self-appraisal regarding commitment to safety.

Objective 4 was to explore whether or not employee scores on the group-level safety climate sub-scale and supervisor self-appraisals were statistically different. If so, the next step would be to determine which are more predictive of safety behavior and injury severity: employee group-level safety climate perceptions or supervisors' self-appraisals (Objective 5).

2. Method

2.1. Participants and procedure

The data used in the current study are a subset of data originally collected as part of a larger study that developed and validated a trucking industry-specific safety climate scale (Huang et al., 2013; Zohar et al., 2014). The details of how the survey was distributed and collected were provided in Huang et al., 2013.

The participants were all long-haul truck drivers. Even though groups of drivers may share the same dispatcher (supervisor), they have little to no opportunity to interact with their co-workers and often do not even know who their co-workers

are. Additionally, the contact between drivers and dispatchers can be sparse, with the main mode of communication being in-vehicle radio devices and cell phones, with many never even meeting in person (Zohar et al., 2014).

The survey was delivered to all the participants using a web-based platform. Data utilized in the study were collected from four trucking companies in the U.S. where both dispatchers and truck drivers participated in the project. The dataset included a matched sample of 1831 employees and their 219 supervisors. The response rates across the companies ranged from 34% to 73%, with a mean around 44%. This is the first time in this series of studies that supervisor data were utilized in the analysis.

Participants were promised confidentiality, using a double-blind coding system approved by the Research Institute IRB. The trucking company did not know who participated and did not have access to the survey data, while the research team did not have the names of the participants. Six months after survey completion, performance outcomes were provided to us from the participating companies. No names were provided; only the company IDs were used to link the information. Supervisor and employee survey results were matched by the project team with performance data. Such a prospective design allows more stringent testing of the predictive validity of the study's variables.

In terms of participants' characteristics, the average age of the 1831 truck drivers was 42.5 (SD = 5.9) years; average tenure as a professional driver was 12.1 (SD = 9.6) years; and average employment at the current company was 5.1 (SD = 5.7) years. All were long-haul drivers and not unionized. Also, 67.9% reported they were paid by the hour and 32.1% of participants were paid by the mile.

For the participating supervisors (n = 219), 22.1% were 20-29 years old, 33.8% were 30-39 years old, 25.0% were 40-49 years old, and 19.1% were over 50. On average, each had been a supervisor for 7.6 years (SD = 6.9) and 35.0% of them had previous experience working as a truck driver. No significant differences were observed with regard to the demographic or descriptive statistics associated with survey variables.

2.2. Measures

2.2.1. Trucking safety climate (TSC)

Safety climate was measured using the 40-item Trucking Safety Climate Scale (Huang et al., 2013). Scale items refer both to company policies and procedures and supervisory practices, accompanied by a 5-point scale ranging from strongly disagree (1) to strongly agree (5). This scale consists of two sub-scales: 20 items of organization-level safety climate (e.g., My company cares more about my safety than on-time delivery; My company turns a blind eye when a supervisor bends some safety rules) and 20 items of group-level safety climate (e.g., Supervisors are strict about driving safely even when we are tired or stressed; My supervisor pushes me to keep driving even when I call in to say I feel too sick or tired). All responses for items with negative wordings were reversely coded. The 20 questions of the organization-level sub-scale were the same for both employees and supervisors, focusing on the company's commitment to and prioritization of safety. For the group-level safety climate sub-scale, there were separate questions for employees and supervisors, such that the referent was the supervisor in both cases. Employee items required them to refer to their supervisor (e.g., ''My supervisor compliments employees who pay special attention to safety'') while supervisors were asked to refer to themselves (e.g., ''I compliment my employees who pay special attention to safety''). Employees were providing an upward-appraisal of their supervisor, while supervisors were providing a self-appraisal regarding commitment to safety.

Internal consistency statistics for employees' organization- and group-level safety climate sub-scales were satisfactory with Cronbach's a = .91 and .94, respectively. For supervisors' organization- and group-level safety climate sub-scales, Cron-bach's a statistics were respectively .90 and .90, also indicating good internal consistency.

2.2.2. Subjective safety outcome measure: safety behavior

The participating truck drivers' self-reported driving safety behavior was measured by six items adapted from Huang, Roetting, McDevitt, Melton, and Smith (2005) and utilized in Huang et al. (2013). An example of the items is ''I always comply with the posted speed limits.'' The items were measured on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree) and demonstrated a marginally acceptable level of inter-item reliability of Cronbach's a = .66.

2.2.3. Objective safety outcome measure: lost work days (injury severity)

Given that the measure of driving safety behavior was based on respondents' self-reports, a socially desirable response tendency could be expected, which might compromise the objectivity of the measure. To address this concern, objective safety data were collected. Road injury, the objective safety criterion used in this study, was operationalized as lost work days (over the past six months) due to injury and was collected six months after the survey implementation. One notable advantage of using the road injury variable is that it can convey comprehensive information about safety outcomes, unlike simple accident or injury frequency, which does not take into account the severity, fatality, or overall cost of incidents. By its nature, the number of lost work days is a count variable, which is less likely to follow a normal distribution (i.e., zero for the majority of truck drivers) and should not be used as a dependent variable in ordinary least-square regression modeling (Cameron & Trivedi, 1998). Thus, a Poisson log-link generalized linear modeling (GLM) approach was used to examine the hypothesized paths from safety climate and safety behavior to lost work days.

2.3. Data analysis procedure

To address Objective 1, measurement equivalence (ME) was tested with a multi-group confirmatory factor analysis (CFA) approach (Raju et al., 2002). CFA is used to verify a hypothesized factor structure of observed variables (i.e., item scores) and test a relationship between observed variables and their underlying latent factors. CFA-based ME testing is carried out by comparing several models with different numbers of equal parameter constraints. Specifically, invariance of measurement structure (i.e., configural invariance), item-loadings onto latent factors (i.e., metric invariance), reference points of item ratings (i.e., scalar invariance), residual variances of items (i.e., residual invariance), variances of latent factors (i.e., factor variance invariance), and covariances of latent factors (i.e., factor covariance invariance) were examined as Vandenberg and Lance (2000) proposed. If these measurement properties are not notably heterogeneous across employees and supervisors, the addition of equal parameter constraints on them would not result in significant model fit worsening, which can be judged by comparative fit index (CFI) decreases greater than .02 (Cheung & Rensvold, 2002) and disappearance of the overlap of root mean square error of approximation (RMSEA) 90% confidence interval (Wang & Russel, 2005). Perfect ME indicates that operations of the safety climate scale's items and latent factors are homogeneous across employees and supervisors. Discrepancy in the safety climate scale score between the two groups when solid ME is established can be attributable to a true score difference in the safety climate, rather than heterogeneous measurement implications.

To address Objectives 2 and 4 (comparison of supervisor versus employee ratings of organization- and group-level safety climate), paired sample t-tests were conducted as well as mixed effect ANOVA. Path analysis was performed to test the links from employee safety climate perception and supervisory interpretation of safety climate to safety behavior and lost work days due to road injury (Objectives 3 and 5). Particularly, a full mediation effect of safety behavior on the safety climate and objective safety outcome relationship was examined. Unlike the original path model, safety climate was divided into two components, such that the safety behavior and objective safety outcome variables were regressed on organization-level safety climate as perceived or interpreted by employees or supervisors.

3. Results

3.1. Preliminary analysis

Tables 1A and B present means, standard deviations, intercorrelations, and reliabilities of study variables for employee and supervisor responses, respectively. Strong correlations were observed between organization- and group-level safety climate dimensions perceived by employees (r = .70, p < .01) and interpreted by supervisors (r = .75, p < .01). Employees' perceptions of organization- and group-level safety climate dimensions were significantly correlated with safe driving behavior with r = .43 and .45 (p < .01), respectively. Lost work days due to road injury showed large deviance statistics (SD = 6.79) with a mean of .33, indicating that the vast majority of participating truck drivers (85.7%) reported zero days lost, as originally expected.

3.2. Organization-level safety climate comparisons

Measurement equivalence (ME) was tested with a multi-group CFA approach. Strong measurement equivalence was observed (Objective 1), showing that appropriate and meaningful comparisons can be made between employee organization-level safety climate and supervisory interpretation of organization-level safety climate. The addition of a series of equal parameter constraints did not result in significant model fit deterioration (i.e., the CFI reductions were not greater than .02 and 90% RMSEA confidence intervals overlapped sequentially for each model). The results can be found in Table 2.

Table 1

Means, standard deviations, and intercorrelations among study variables.

Mean (SD)

A. Employees (n = 1831)

1.OSC 3.95 (.68) (.91)

2. GSC 4.05 (.76) .75** (.94)

3. Safety behavior 4.39 (.60) .43** .45** (.66)

4. Lost work days .33 (6.79) -.03 -.02 -.07*'

Mean (SD) 1 2

B. Supervisors (n = 219)

1. OSC 4.18 (.57) (.90)

2. GSC 4.45 (.48) .70** (.90)

Notes: OSC: organization-level safety climate, GSC: group-level safety climate. Injury severity variable is a count variable and its correlations with other variables should not be interpreted as the ordinary zero-order correlation.

** p <.01

A paired-sample t-test showed that supervisory interpretation of organization-level safety climate and employee organization-level safety climate scores were significantly different, t(1830) = -6.41, p < .01 (Objective 2). The results of this analysis are located in Table 3. Specifically, results showed that scores of supervisors' interpretations were statistically higher than those of their employees. In the paired-sample data, employee perception and supervisor ratings were significantly, but weakly correlated (r = .05, p < .05). In order to supplement the limitation of applying paired sample t-test to the nested data in which numbers of employees and supervisors are not equal (i.e., measures from the same supervisor are not independent), a mixed effect ANOVA was conducted. This method separates the within- and between-group variability to resolve the dependency problem associated with a nested data structure (Raudenbush & Bryk, 2002). The result was fairly similar to the paired t-test with y00 = -.16 (SE = .05, p < .01), indicating significantly higher scores for supervisory interpretation of safety climate than employee safety climate perceptions. Path analysis (Fig. 1) showed that employees' organization-level safety climate better predicted safety behavior (directly) and injuries (indirectly) than did supervisors' interpretations (Objective 3). Specifically, employee organization-level safety climate was associated with safety behavior with an unstan-dardized coefficient = .36 (SE = .05, p < .01), while supervisor interpretations were not significantly related to safety behavior with an unstandardized coefficient of .03 (SE = .05, p = .59). Safety behavior was linked with the number of lost work days due to injury with an unstandardized coefficient of -1.24 (SE = .17, p < .01). The full mediation effect of safety behavior in linking the relationship between employee organization-level safety climate and lost work days was -.45 (SE = .09, Sobel's Z = -7.09, 20,000 bootstrapped 95% CI = -.63 to -.29) and it was statistically significant (p < .01). However, the full mediation effect of safety behavior in linking the relationship between supervisory interpretation of organization-level safety climate and lost work days was not significant (indirect effect = -.03, SE = .06, Sobel's Z = -.54, 20,000 bootstrapped 95% CI = -.15 to .08). Together, employee perception and supervisory interpretation of organization-level safety climate explained 17% of the total safety behavior variance.

3.3. Group-level safety climate comparisons

A paired-sample t-test showed that employee scores on the group-level safety climate sub-scale and supervisor self-appraisals were statistically different, t(1810) = -22.73, p <.01 (Objective 4). The results of this analysis are located in Table 4.

Specifically, supervisors' self-appraisals of their own commitment and prioritization of safety were higher than their employees' perceptions. Employees' and supervisors' ratings in the paired-sample data were significantly, but weakly correlated (r = .06, p < .01). The mixed effect ANOVA result (y00 = -.47, SE = .04, p < .01) was congruent with the paired t-test.

In the path analysis (Fig. 2), employees' group-level safety climate predicted safety behavior (directly) and injuries (indirectly) better than supervisors' interpretations (Objective 5). Group-level safety climate perceived by employees was linked to safety behavior with an unstandardized coefficient of .33 (SE = .05, p < .01), while supervisors' rating was only minimally linked to safety behavior with an unstandardized coefficient of -.01 (SE = .04, p = .63). The relationship between safety behavior and the number of lost work days due to injury was significant (unstandardized coefficient = -1.24, SE = .17, p < .01). The relationship between employees' group-level safety climate and lost work days was fully mediated by safety behavior with an unstandardized coefficient of -.40 (SE = .09, Sobel's Z = -4.73, 20,000 bootstrapped 95% confidence interval = -.58 to -.25), which was statistically significant (p < .01). However, the mediating effect of safety behavior on the association between supervisors' rating and lost work days was not significant (coefficient = .02, SE = .05, Sobel's Z = .33, 20,000 bootstrapped 95% CI = -.08 to .11). Employee group-level safety climate perceptions and supervisor interpretations could explain 17% of the total safety behavior variance.

Table 2

Measurement equivalence testing with multi-group CFA approach (1831 employees & 219 supervisors).

TSC-O V2 (df) CFI TLI RMSEA (90% CI)

Step 1: configural invariance model 2109.229 (334) .987 .983 .051 (.049-.053)

Step 2: metric invariance model 2206.092 (351) .986 .983 .051 (.049-.053)

Step 3: scalar invariance model 2587.162 (371) .983 .981 .054 (.052-.056)

Step 4: residual invariance model 2992.799 (391) .980 .979 .057 (.055-.059)

Step 5: factor variance invariance model 3025.060 (394) .980 .979 .057 (.055-.059)

Step 6: factor co-variance invariance model 3039.589 (397) .980 .979 .057 (.055-.059)

Notes: TSC-O = Trucking Safety Climate - Organization Level.

CFI: comparative fit index. TLI: Tucker-Lewis index. RMSEA: root mean square error of approximation. CI: confidence interval.

Step 1: configurai invariance model (equal factor structure).

Step 2: metric invariance model (equal factor structure + equal factor loadings).

Step 3: scalar invariance model (equal factor structure + equal factor loadings + equal intercepts).

Step 4: invariant uniqueness model (equal factor structure + equal factor loadings + equal intercepts + equal uniquenesses).

Step 5: invariant factor variance model (equal factor structure + equal factor loadings + equal intercepts + equal uniquenesses + equal factor variance). Step 6: invariant factor covariance model (equal factor structure + equal factor loadings + equal intercepts + equal uniquenesses + equal factor variance + equal factor covariance).

Table 3

Paired-sample t-test comparing employee organization-level safety climate perceptions and supervisory interpretations (1831 employees & 219 supervisors).

Paired Differences t df

Mean S.E. 95% CI

TSC-O -.13 .02 -.17 to -.09 -6.41" 1830

Notes: TSC-O = Trucking Safety Climate - Organization Level.

Fig. 1. Indirect effect of organization-level safety climate perceptions on lost work days due to road injury via safety behavior. Note: TSC-O-emp = employee organization-level trucking safety climate perceptions; TSC-O-sup = supervisory interpretation of organization-level trucking safety climate; ** p < .01.

Table 4

Paired-sample t-test comparing employee group-level safety climate perceptions and supervisory interpretations (1831 employees & 219 supervisors).

Paired differences t df

Mean S.E. 95% CI

TSC-G -.46 .02 -.5 to -.42 -22.73" 1810

Notes: TSC-G = Trucking Safety Climate-Group-level.

Fig. 2. Indirect effect of group-level safety climate on lost work days due to road injury via safety behavior. Note: TSC-G-emp = employee group-level trucking safety climate perceptions; TSC-G-sup = supervisory interpretation of group-level trucking safety climate; ** p < .01.

4. Discussion

This study increases our understanding of the lone worker context by examining employee safety climate and supervisory interpretations of safety climate, how similar or different they are, and how they are related to important safety outcomes.

Results from Objective 1 provided support for the measurement equivalence of the Trucking Safety Climate Scale at the organization level for both employees and supervisors. This means that, regardless of whether respondents were employees or supervisors, there was an ''equality of item level and sub-scale level true scores for persons with identical latent scores'' (Raju et al., 2002). This finding highlights the utility of the scale, as it shows that it can be used to make meaningful comparisons between employee perceptions and supervisory interpretations of organization-level safety climate. The same method can be generalized to other research questions when data from different groups are collected (e.g., when conducting cross-cultural comparisons by collecting data from people in different cultures/countries) using the same scale.

After measurement equivalence was established for organization-level safety climate, the results of the study revealed that supervisors' and employees' scores were statistically different and supervisors' interpretations of both organization-and group-level safety climate were higher than employees' perceptions (as shown from the results of Objectives 2 and 4). Moreover, for both organization- and group-level safety climate scores, only the employees' perceptions were predictive of safety outcomes (as shown from the results of Objectives 3 and 5). This is a striking result; although supervisors and employees were interpreting the scales in the same way (as assessed by measurement equivalence), their scores were significantly different. This finding occurs despite the fact that lone workers' primary source of information, upon which their safety climate perceptions were based, was their supervisors (e.g., Zohar et al., 2014). Ironically, the employees' (supervisor-informed) safety climate perceptions were more predictive of safety outcomes than the safety climate interpretations of the supervisors themselves.

As stated earlier, we have utilized prior research to propose separate theoretical explanations for supervisor-employee differences at the organization level and group level. However, given the observed pattern of results, we believe that it would be more useful to propose a more integrated argument which addresses organization and group levels within a single framework. This framework includes the following assumptions: (a) For lone workers, unlike in traditional workplaces, a primary

source of information upon which employee safety climate perceptions are based, both organizational- and group-level, is likely to be the actions of their supervisor; and (b) The knowledge structures contained within the supervisors' mental models provide the foundation for our theoretical proposals.

Our integrated approach includes elements of the following: gatekeeper model (Zohar & Luria, 2010), increased opportunity to observe approach (Lawler, 1967), social desirability (Maccoby & Maccoby, 1954), and self-assessment bias (Mount, 1984a,b; Borman, 1997). For the gatekeeper model there are two likely patterns which are characteristic of supervisor behavior. In the first pattern, the supervisor correctly assesses that top management's policy can be characterized by parity between enacted and espoused supports of safety. Being closer to top management compared to the lone workers, supervisors may be better able to discern the true priorities of upper management. However, in their gatekeeper role, supervisors bring in their own values when interpreting company climate to their group members. For example, a supervisor may be willing to disregard safety policies/procedures when production falls behind and expect drivers to respond to cell phone communications when on the road. Consider how this situation may be represented in this supervisor's mental model. The supervisor may internalize an accurate representation of top management policy in the declarative knowledge structure component of his mental model. At the same time, the influence of social desirability and self-assessment biases lead to a representation of his/her own interpretation of management policy as also positive, and close to that of management. In contrast, the actual operational decisions of the supervisor (e.g., calling drivers on their cell phones) are based on the procedural components of his/her mental model and form the primary source of information upon which his/her employees base their safety climate perceptions. This is manifest in the distinction between sent and received messages (Katz & Kahn, 1978; Mer-ton, 1957; Zohar & Polachek, 2014). Interestingly, it is in the content of the received messages that the predictive power of employee ratings of safety climate reside.

Thus, in responding to the organizational-level safety climate survey questions, the supervisor will rely on the relatively more accurate set of declarative knowledge structures relating to top management internalized in his/her mental model. However, in responding to the group-level self-assessment safety climate questions, he/she will rely on knowledge structures relating to his/her own interpretation of management policy; structures likely subject to social desirability and self-serving bias. In contrast, employees will utilize the information originating in supervisor behaviors derived from procedural knowledge structures in the supervisor mental model to respond to both the organization-level and group-level surveys. Consequently, employee scores on both surveys are significantly lower than supervisor scores, and are predictive of safety outcomes. Supervisor scores, on the other hand, are based on information which is discrepant with the actual situation on the ground and do not predict safety outcomes.

Further evidence for this explanation rests on the fact that the correlation between employees' organizational- and group-level safety climate scores is very high (r = 0.75). This would support the assertion that the primary source of information upon which employees base their organizational safety climate perceptions is the supervisor.

This explanation accounts for only one aspect of the gatekeeper model. In the 2nd pattern, one might expect some supervisory gate-keepers to be more supportive of their drivers than top management. For example, top management may not set a specific policy on cell phone use, but supervisors communicate to drivers the importance of not using cell phones while driving and they never call them when they are on the road, behaving in accordance with this viewpoint. Under this situation, it would be expected that supervisors would interpret organizational safety climate as lower than employees perceive it to be. We did not find evidence to support this pattern of behavior. As it happens, out of the eight companies in the overall study (Huang et al., 2013), the four companies that (a) agreed to allow supervisors to participate and (b) provided objective safety outcomes, also had the highest average safety climate scores. It is perhaps the case that, since all the supervisors in the current study came from companies with high overall safety climate scores, it is more likely that these supervisors may reflect the first gatekeeper pattern mentioned above in which the supervisors are less supportive than top management. An obvious way to more clearly test this explanation would be to replicate the investigation with a sample that contained companies across a broader range of safety climate.

In the conduct of this study, attempts were made to minimize social desirability bias (Maccoby & Maccoby, 1954). Participants were assured of confidentiality of their responses, web surveys were used which both supervisors and drivers could answer privately, a double-blind coding system was used to identify participants so that no names were collected, and participants were informed that only group-level aggregated data would be provided back to the company. The relative accuracy of the employee's perceptions at both the organizational- and group-level, as evidenced by significant predictions of safety outcomes, would argue that social desirability influences, if present, were less evident for the workers than for supervisors. For the organizational-level scores, the data suggest that supervisors were more likely to interpret management policies in a positive direction.

As discussed earlier, without additional qualitative research, the actual reasons causing the differences between supervisors and employees' perceptions in the study are uncertain. Future research should further investigate the root causes of these perceptual gaps and take the root causes into account when designing and implementing interventions to improve overall workplace safety.

4.1. Practical Implications

The results have important practical implications. When supervisors' interpretations of safety climate are higher than those of their employees, it means they perceive that the company prioritizes safety more strongly than the employees

do. These differences could lead to management failing to make necessary changes regarding the company's safety policies, procedures, and practices if only management's perceptions are used to make decisions. If the supervisors' scores for organization-level safety climate do, in fact, reflect true priorities of upper management (i.e., prioritizing safety over competing demands), it is important that this information be passed down to front-line employees because supervisor interpretations alone are not directly related to employee safety outcomes. This may present a challenge to organizations in the trucking industry (and other organizations employing lone workers), as their front-line workers have limited opportunities to interact with top management.

The results of the study further revealed that supervisors' interpretations of safety climate were not leading indicators of employee future safety outcomes; only employee perceptions significantly predicted safety behaviors and future injuries. The findings of this study suggest that employees' safety climate scores can be used to help managers and supervisors be more realistic in their understanding of the way in which the company's values regarding safety are perceived by their workers (e.g., Cheyne et al., 2003). Using a safety climate scale to understand their employees' points of view may help supervisors identify issues and make subsequent changes that will improve employees' safety climate perceptions, thus decreasing accidents and injuries.

Another practical implication is the importance of utilizing employee perceptions to inform potential training and interventions. Sometimes information about the current status of particular jobs and work is obtained from supervisors rather than employees; however, it may be the case that supervisors are not always aware of the decisions that workers face between working harder or faster and heeding all possible safety hazards. Our results suggest that there is a need to improve the correspondence between supervisors' safety climate interpretations and employees' safety climate. Communication with one's supervisor is still one of the main ways employees are informed about organizational policies and procedures and the types of behaviors that are rewarded or punished (Zohar et al., 2014), especially in the case of lone workers. Literature has shown that supervisor training on increasing safety communication with their employees can improve employees' safety outcomes (Zohar & Luria, 2003; Zohar & Polachek, 2014). In this case, providing safety training on communication or more opportunities for two-way communication between supervisors and employees may be used to bridge the gap. Again, this study points to the necessity of considering employee perceptions, as they are more closely related to the outcomes of interest (at least in the case of lone workers).

4.2. Limitations

Overall, the contributions of this study should be viewed in light of several limitations. First, although objective outcome data were collected six months after survey administration, the survey itself was cross-sectional. A future longitudinal study will provide stronger evidence for causal relationships between safety climate, safety behavior and outcomes. Second, this study focuses on the results of safety climate surveys administered to employees and supervisors in a lone working situation specific to the trucking industry. This research could also be extended to lone workers in other job types and industries. Third, this study utilized only one objective outcome (lost work days). Future research would benefit from utilizing other outcomes (e.g., accident near-misses), as well as data collected over a longer period of time (for example, one year after survey administration). Furthermore, similar to other survey studies, the data in this study were self-reported (except for the objective outcome of injury severity) and social desirability effects (and/or self-assessment biases from supervisors) might have been a factor. Caution needs to be used when interpreting the validity of these survey results. Lastly, additional studies should compare safety climate interpretations and perceptions of other employees (in addition to front-line workers and their direct supervisors), including safety managers and upper management.

5. Conclusion

The results of this study make significant contributions to the safety climate literature. First, the study demonstrates measurement equivalence of the organization-level safety climate sub-scale items among supervisor interpretations of safety climate and employee safety climate. Second, it provides empirical evidence that supervisor interpretations of safety climate are not congruent with employee perceptions for lone workers, and supervisors give higher ratings. Third, employee safety climate perceptions, but not supervisor interpretations of safety climate, significantly predicted safety behavior and days away from work due to injury (an indicator of injury severity). The results provide support for traditional safety climate literature, and suggest that when trying to gauge and improve upon a trucking company's safety climate, employees' perceptions are more indicative of safety behaviors and outcomes than supervisors' interpretations.

Acknowledgements

The authors wish to thank the following team members for their invaluable assistance: Dov Zohar (Technion - Israel Institute of Technology), Mo Wang (University of Florida), Garry Gray (Harvard School of Public Health), Susan Jeffries, Peg Rothwell, Anna McFadden and Ryan Powell (Liberty Mutual Research Institute for Safety) for data collection, analysis and general assistance; Dave Melton, Dave Money, Jim Houlihan (Liberty Mutual Insurance), and Keith Herzig (Herzig Hauling) for technical consulting.

References

Argyris, C., & Schon, D. A. (1996). Organizational learning II: Theory, method, and practice. Reading, MA: Addison- Wesley.

Baruch, Y., & Nicholson, N. (1997). Home, sweet work: Requirements for effective home working. Journal of General Management, 23,15-30.

Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. Psychological Bulletin, 117, 497-529.

Beus, J. M., Payne, S. C., Bergman, M. E., & Arthur, W. R. (2010). Safety climate and injuries: An examination of theoretical and empirical relationships. Journal of Applied Psychology, 95, 713-727.

Blickensdorfer, E., Cannon-Bowers, J. A., Salas, E., & Baker, D. P. (2000). Analyzing knowledge requirements in team tasks. In J. M. Schraagen, S. F. Chipman, & V. L. Shalin (Eds.), Cognitive task analysis. Mahway, N.J.: Lawrence Earlbaum, Associates.

Borman, W. C. (1997). 360 degree ratings: An analysis of assumptions and a research agenda for evaluating their validity. Human Resource Management Review, 7, 299-315.

Bureau of Labor Statistics. (2012). Census of fatal occupational injuries (CFOI) - Current and revised data. <http://www.bls.gov/iif/oshcfoi1.htm>.

British Security Industry Association. (2010). Lone workers guide for employers. <http://www.bsia.co.uk/publications.php?r=O4R7JX948641> Retrieved November 2012.

Cameron, A. C., & Trivedi, P. K. (1998). Regression analysis of count data. Cambridge University Press.

Cheung, G. W., & Rensvold, R. B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233-255.

Cheyne, A., Tomas, J. M., Cox, S., & Oliver, A. (2003). Perceptions of safety climate at different employment levels. Work & Stress: An International Journal of Work, Health & Organisations, 17(1), 21-37.

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, 94,1103-1127.

Cox, S., Tomas, J. M., Cheyne, A., & Oliver, A. (1998). Safety culture: The prediction of commitment to safety in the manufacturing industry. British Journal of management, 9, 3-11.

Eakin, J. M. (1992). Leaving it up to the workers: sociological perspective on the management of health and safety in small workplaces. International Journal of Health Services, 22, 689-704.

Gajendran, R. S., & Harrison, D. A. (2007). The good, the bad, and the unknown about telecommuting: Meta-analysis of psychological mediators and individual consequences. Journal of Applied Psychology, 92,1524-1541.

Golden, T. D., Veiga, J. F., & Dino, R. N. (2008). The impact of professional isolation on teleworker job performance and turnover intentions: does time spent teleworking, interacting face-to-face, or having access to communication-enhancing technology matter? Journal of Applied Psychology, 93,1412-1421.

Graen, G. B., & Uhl-Bien, M. (1995). Relationship-based approach to leadership: Development of LMX theory of leadership over 25 years. Leadership Quarterly, 6, 219-247.

Griffin, 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, 347-358.

Huang, Y. H., Zohar, D., Robertson, M. M., Garabet, A., Lee, J., & Murphy, L. A. (2013). Development and validation of safety climate scales for lone workers using truck drivers as exemplar. Transportation Research Part F: Traffic Psychology and Behavior, 17, 5-19.

Huang, Y. H., Chen, P. Y., & Grosch, J. W. (2010). Safety climate: New developments in conceptualization, theory, and research. Accident Analysis and Prevention, 42, 1421-1422.

Huang, Y. H., Roetting, M., McDevitt, J. R., Melton, D., & Smith, G. S. (2005). Feedback by technology: Attitudes and opinions of truck drivers. Transportation Research Part F, 8, 277-297.

Huang, Y. H., Verma, S. K., Chang, W. R., Courtney, T. K., Lombardi, D. A., Brennan, M. J., et al (2012). Supervisor vs. employee safety perceptions and association with future injury in US limited-service restaurant workers. Accident Analysis and Prevention, 47, 45-51.

Hughes, P., & Ferrett, E. (2009). Introduction to health and safety at work. Oxford: Butterworth Heinemann, p. 117.

Katz, D., & Kahn, R. L. (1978). The social psychology of organizations, 2nd ed., New York: Wiley.

Lawler, E. E. (1967). The multitrait-multirater approach to measuring managerial job performance. Journal of Applied Psychology, 51, 369-381.

Leigh, J. P. (2011). Economic Burden of occupational injury and illness in the United States. Milbank Quarterly, 89, 728-772.

Lewin, K., Lippitt, R., & White, R. K. (1939). Patterns of aggressive behaviour in experimentally created social climates. Journal of Social Psychology, 10, 271-301.

Liberty Mutual Research Institute for Safety, (2012). 2012 Liberty mutual workplace safety index. Hopkinton, MA. <www.libertymutualgroup.com/research>.

Mount, M. K. (1984a). Supervisor, self-, and subordinate ratings of performance and satisfaction with supervision. Journal of Management, 10, 305-320.

Maccoby, E. E., & Maccoby, N. (1954). The interview: A tool of social science. In G. Lindzey (Ed.), Handbook of Social Psychology (1st ed., pp. 449-487). Cambridge, MA: Addison-Wesley.

Merton, R. K. (1957). Continuities in the theory of reference groups and social structure. In R. K. Merton (Ed.), Social theory and social structure (pp. 281-386). Glencoe, IL: Free Press.

Mount, M. K. (1984b). Psychometric properties of subordinate ratings of managerial performance. Personnel Psychology, 37, 687-702.

National Health Service (NHS), (2005). Not alone: A guidance for lone workers. <http://www.nhsbsa.nhs.uk/SecurityManagement/Documents/ Lone_Working_Guidance_final.pdf. Retrieved October 2012>.

Paté-Cornell, M. E. (1990). Organizational aspects of engineering system. Safety Science, 250,1210-1216.

Raju, N. S., Laffitte, L. J., & Byrne, B. M. (2002). Measurement equivalence: A comparison of methods based on confirmatory factor analysis and item response theory. Journal of Applied Psychology, 87, 517-529.

Reason, J. T. (1998). Achieving a safe culture: theory and practice. Work and Stress, 12, 293-306.

Raudenbush, S. W., & Bryk, A. S. (2002). Hierarchical linear models: Applications and data analysis methods (2nd ed.). Newbury Park, CA: Sage.

Rouse, W. B., & Morris, N. M. (1986). On looking into the black box: Prospects and limits in the search for mental models. Psychological Bulletin, 100, 349-363.

Simons, T. L. (2002). Behavioral integrity: The perceived alignment between managers' words and deeds as a research focus. Organization Science, 13,18-35.

Stout, R. J., Cannon-Bowers, J. A., & Salas, E. (1996). The role of shared mental models in developing team situational awareness: Implications for training. Training Research Journal, 2, 85-116.

Thompson, J. W. (1996). Employee attitudes, organizational performance, and qualitative factors underlying success. Journal of Business and Psychology, 11, 171-196.

Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance literature Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3, 4-70.

Wright, C. (1986). Routine deaths: Fatal accidents in the oil industry. The Sociological Review, 34, 265-289.

Wang, M., & Russel, S. S. (2005). Measurement equivalence of the job descriptive index across Chinese and American workers: Results from confirmatory factor analysis and item response theory. Educational and Psychological Measurement, 65, 709-732.

Zohar, D. (1980). Safety climate in industrial organizations: Theoretical and applied implications. Journal of Applied Psychology, 65, 96-102.

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, 587-596.

Zohar, D. (2008). Safety climate and beyond: A multi-level multi-climate framework. Safety Science, 46, 376-387.

Zohar, D. (2010). Thirty years of safety climate research: reflections and future directions. Accident, Analysis and Prevention, 42(5), 1517-1522.

Zohar, D. (2011). Safety climate: Conceptual and measurement issues. In J. C. Quick & L. E. Tetrick (Eds.), Handbook of occupational health psychology (2nd ed., pp. 141-164). Washington, DC: American Psychological Association.

Zohar, D., Huang, Y. H., Lee, J., & Robertson, M. M. (2014). A mediation model linking dispatcher leadership and work ownership with safety climate as predictors of truck driver safety performance. Accident Analysis and Prevention, 62,17-25.

Zohar, D., & Luria, G. (2003). The use of supervisory practices as leverage to improve safety behavior: A cross-level intervention model. Journal of Safety Research, 34, 567-577.

Zohar, D., & Luria, G. (2010). Group leaders as gatekeepers: Testing safety climate variations across levels of analysis. Applied Psychology: An International Review, 59(4), 647-673.

Zohar, D., & Polachek, T. (2014). Discourse-based intervention for modifying supervisory communication as leverage for safety climate and performance improvement: a randomized field study. Journal of Applied Psychology, 99(1), 113-124.

Supervisory interpretation of safety climate versus employee safety climate perception: Association with safety behavior and outcomes for lone workers Academic research paper on "Economics and business"

Abstract of research paper on Economics and business, author of scientific article — Yueng-hsiang Huang, Michelle M. Robertson, Jin Lee, Jenn Rineer, Lauren A. Murphy, et al.

Similar topics of scientific paper in Economics and business , author of scholarly article — Yueng-hsiang Huang, Michelle M. Robertson, Jin Lee, Jenn Rineer, Lauren A. Murphy, et al.

Academic research paper on topic "Supervisory interpretation of safety climate versus employee safety climate perception: Association with safety behavior and outcomes for lone workers"