**1. Introduction**

Accidents at work have been considered as an important topic due to their profound impact on human life at the physical, mental, and economic levels. Accidents at work function as a fatal blow not only to the victims, but also to their firms and the national economy. The mental and physical damages caused by accidents at work usually make the daily life of the victims devastating for a lifetime. Furthermore, companies with accidents at work are more likely to suffer serious economic losses and long recovery periods. Therefore, efforts to reduce accidents at work are highly required to protect employees' well-being as well as firms' sustainability.

According to Bird and his colleague's domino theory [1], before the occurrence of an accident, a "precursor" appears which functions as the direct cause of the accident. Among various precursors, employees' unsafe behavior (e.g., impulsive or careless behavior) has been regarded as one of the most direct antecedents of an accident. To enhance the quality of employees' safety behavior, previous studies have suggested various factors influencing safety behavior. For example, the job demand–resources (JD-R) model [2,3] suggests a useful conceptual model which describes the impact of various physical, psychological, and organizational factors on safety outcomes. The model points out that two types of working conditions including job demands and job resources significantly affect safety outcomes such as safety behavior, accidents, and injuries. The job demands dimension consists of various components that are negatively associated with safety behavior, including a bad physical environment, a high level of work pressure, complexities, and risks. In addition, the job resources dimension consists of various factors that are positively related to safety behavior, including knowledge, autonomy, and a supportive environment (e.g., social support, leadership, a safety climate).

Among the various antecedents, in this paper, we focus on leadership styles, especially transformational leadership, due to its critical role in explaining various organizational outcomes by affecting employees' cognitions, emotions, and behavior [4–6]. Transformational leadership is defined as a leadership style "broadening and elevating followers' goals and providing them with confidence to perform beyond the expectations specified in the implicit or explicit exchange agreement" [6]. Although some scholars have investigated the relationship between transformational leadership and employees' safe behavior [7–9], we believe that there are some research gaps to be additionally addressed.

First, the previous studies which delved into the association between transformational leadership and safety behavior have underexplored the importance of job strain in explaining the intermediating mechanism of the relationship. Job strain can be defined as an employee's negative perceptions, emotions, and physiological states which emerge when an employee recognizes that he or she cannot adequately deal with various external stimuli such as interpersonal relationships, job characteristics, and the work environment [10–13]. The concept has been known to critically influence employees' perceptions, attitudes, and behaviors by diminishing cognitive/emotional/physical abilities to implement his or her tasks and duties. Then, eventually, it deteriorates the quality of his or her performance in an organization [11–13]. In other words, job strain of employees functions as a critical construct which not only explains the influence of transformational leadership on employees in an organization, but also predicts their important attitudes or behaviors. Despite its significant theoretical and practical impact on employees in an organization, to the best of our knowledge, there has not been any research to have considered job strain as a critical mediator in explaining the influence of transformational leadership on employees' safe behavior. Although some previous studies have suggested a bivariate relationship between transformational leadership and job strain [7] as well as between job strain and safety behavior [12], those did not examine the entire relationship among the three variables in an integrated manner based on an overarching theoretical ground. By investigating the intermediating role of job strain with a theoretically overarching basis, we can provide an elaborate explanation on how transformational leadership positively affects safety behavior at work. Thus, examining the role of job strain in describing the association is highly required.

Second, previous studies on the transformational leadership–safety behavior link have paid less attention to the contingent or contextual factors that moderate the relationship. These studies have focused on various intermediators in the link. For example, Shen and his colleagues [9] reported that transformational leadership influences employees' safety behavior through the sequential mediating roles of safety-specific leader-member exchange (LMX), safety climate, safety knowledge, and safety motivation. Although finding various mediating factors in the relationship between transformational leadership and safety behavior is very important, it is not enough to elaborately describe the influence of transformational leadership on safety behavior because the mediating variables cannot fully explain the association in all situations or contexts. In other words, the explanation of the mediators is limited to a certain situation or context. Thus, we suggest that examining contingent variables (i.e., moderators) in the link would contribute to elaborating the transformational leadership–safety behavior literature.

To deal with these issues, in the present study, we have investigated the intermediating effect of job strain between transformational leadership and safety behavior, as well as the moderating effect of self-efficacy regarding safety on the job strain–safety behavior link. The theoretical logic is based on the context-attitude-behavior model [14,15], which explains the mediation structure. Grounded on it, we expect that transformational leadership, as one of the important contexts, may build employees' behavior (i.e., safety behavior) by affecting their attitude (i.e., job strain). Based on previous works, we suggest that transformational leadership would decrease the level of employees' job strain [7,16,17] and that employees' job strain would increase safety behavior [12,18–21]. In the present study, considering that stress indicates changes in well-being because of various stressors while strain means lowered levels of well-being or functioning (e.g., exhaustion, tension, anxiety, and rumination) [11,12,17], we focus on strain of employees at work.

In addition, we propose that there may be contingent factors that moderate the relationship between job strain and safety behavior. Although job strain may diminish the quality of employees' safety behavior, some buffering factors can weaken the negative influence of job strain on safety behavior. Among those factors, we have focused on employees' self-efficacy regarding safety since the concept of self-efficacy has been regarded as one of the most fundamental factors which explain an individual's perceptions, attitudes, and behavior [22,23]. Based on Eden and Zuk's definition [24], we have defined self-efficacy regarding safety as an individual's overall estimate or expectation of his or her ability to effectively deal with safety-related situations.

When an employee has a high level of self-efficacy regarding safety, he or she can protect himself/herself from the harmful influences of job strain. Furthermore, the employee may feel that he or she has enough abilities to effectively deal with the harmful effects of job strain at work. Through this, the negative psychological and physical states from job strain would not substantially diminish the quality of safety behavior. On the contrary, if the level of an employee's self-efficacy regarding safety is low, he or she may feel that he or she cannot adequately deal with many issues from unsafe situations at work. Then, the negative influences of job strain may be facilitated and amplified, being considerably damaging to his or her safe behavior.

To empirically test the above hypotheses, by utilizing data from 997 employees in South Korea, we conducted a structural equation modeling (SEM) analysis with a moderated mediation model.

#### **2. Theories and Hypotheses**

#### *2.1. Transformational Leadership and Job Strain*

Some previous works on leadership have shown that transformational leadership decreases the degree of an employee's job strain [7,16,17]. According to Cummings and Cooper [25], stress can be defined as "the force that causes a strain in the physical and psychological states by escaping from the state of stability". As the concept of stress develops, there have been a lot of discussions about the concept as well as attempts to describe it in terms of industrial/organizational psychology and organizational behavior theories [11,12]. In the present study, considering that stress indicates changes in well-being because of various stressors while strain means lowered levels of well-being or functioning (e.g., exhaustion, tension, anxiety, and rumination) [11,12,17], we have focused on strain of employees at work. An employee's job strain has been considered as an important factor which diminishes cognitive/emotional/physical abilities to implement his or her tasks and duties, eventually deteriorating the quality of his or her performance [10–13]. Based on many previous works on job strain, we have defined job strain as an employee's negative perceptions, emotions, and physiological states which emerge from recognizing that he or she cannot adequately deal with various external stimuli such as interpersonal relationships, job characteristics, or the work environment [11–13,26,27].

Through providing employees with a higher level of inspirational motivation, idealized influence, supportive caring, and intellectual challenges, transformational leadership can reduce various uncomfortable perceptions and emotions in an organization [7,16,17]. Specifically, we suggest that each of the sub-dimensions of transformational leadership may decrease the level of employees' job strain as follows.

First, a leader with transformational leadership is likely to provide inspirational motivation to his or her followers. Through this, the followers of the leader may pursue noble values and goals beyond their own ego-centric interests [28,29]. Then, the followers may try to proactively cooperate with their colleagues to achieve novel objectives, which tend to be collective-level. Through this collaborative atmosphere, the followers may feel that they are safe and trusted by colleagues in the organization. Those positive perceptions would decrease their level of job strain.

Second, through idealized influence, a leader with transformational leadership can provide psychological safety to his or her followers, directly contributing to reducing their job strain [28]. The followers of transformational leadership are likely to regard the leaders as their role model, identifying with their leader [30]. Through this identification, the followers would feel that they are competent enough to achieve collective and noble goals like their transformational leader. Their positive perceptions and feelings would alleviate their job strain.

Third, a transformational leader tends to stimulate employees to think differently with traditional practices. Through this intellectual stimulation, the followers may try to solve various work problems with novel approaches [30]. The new approaches would help the followers to reframe their stressful experiences in positive and effective ways [31]. Furthermore, in the process of stimulation, followers may effectively solve difficult problems in an organization. Their positive experiences would enhance a follower's sense of self-efficacy, facilitating the more efficient problem-solving ability of followers. By virtue of it, the level of employees' job strain would be decreased.

Lastly, by providing individual consideration, a transformational leader is likely to reduce employees' job strain. It is self-explanatory that the leadership style can decrease the level of employees' strain since the leader's supportive caring for followers in an individual manner would function as a psychological base for them. The followers may feel that their negative emotions such as anxiety, fatigue, and anger in the process of working are healed and restored, by relying on the psychological base [5,28]. Based on the reasons described above, we suggest that transformational leadership is negatively associated with employees' job strain in an organization.

### **Hypothesis 1.** *Transformational leadership is negatively related to employees' job strain.*
