The Effect of Safety Attitudes on Coal Miners’ Human Errors: A Moderated Mediation Model

Abstract

With the advancement of digital technology, the operation scenarios and work of miners have changed. Although the rate of accidents caused by unsafe conditions is decreasing year by year, the rate of accidents caused by human errors is still high. To investigate the influencing factors of miners’ human errors in the context of smart mines, based on the attitude–behavior model, this paper introduced two variables, situational awareness and task complexity, and established a moderated mediation model to explore the path of safety attitudes on human errors. Using time-lagged data from 246 full-time miners working at smart mines, we found that miners’ safety attitudes were effective in reducing human errors, miners’ safety attitudes reduced human errors through the mediation of situational awareness, and task complexity enhanced the positive relationship between safety attitudes and situational awareness, thus positively moderating the indirect relationship between safety attitudes and human errors. The findings advance a new understanding of how safety attitudes can reduce miners’ human errors. They also provide practical implications on the importance of leadership and finding ways to promote situational awareness as well as maintaining good safety attitudes.

1. Introduction

Currently, with the development of smart mines, there is an important shift in the role of human factors in the new mine system [1,2]. The cognitive processes and behavioral responses of workers have fundamentally changed as mine workers have gradually changed from manual manipulation to integrated computer control [3,4,5]. In the new production model, “human”, as the only dynamic link in the “human–machine–environment” system, is always the core element and is the key to connecting all aspects of the entire safety production system [6,7]. Studies have shown that human errors are still the main cause of accidents, and its proportion in mine accidents is over 80% and has remained consistently high over the years [8]. Therefore, in this context, an in-depth investigation of human errors in smart mining systems is of great value to reduce accidents and injuries brought about by the new production model.

In recent years, domestic and foreign scholars have focused on three aspects: the human–machine interface, organizational factors, and personnel characteristics in terms of the factors influencing human errors [9,10,11,12,13,14,15]. In terms of the human–machine interface, Minyan Xia pointed out that to reduce the occurrence of human-caused incidents, it is necessary to develop and design a simple and easy-to-learn human–machine interface so that operators can operate at an experience-based level [9]; in terms of organizational factors, Vinodkumar examined how organizational factors affect human accidents at four levels: inadequate organizational culture, organizational normative failure, organizational communication failure, and organizational functional failure [10]. Chen et al. showed that leaders’ safety attitudes and safety leadership can also have an impact on employees in the organization [11]; in terms of personnel characteristics, most domestic and international studies have analyzed the influence of human personality, energy, knowledge, and experience on unsafe behavior from the perspective of psychology and behavioral disciplines [12,13]. For example, Kao found on this basis that the level of individual safety knowledge causes significant differences in individual safety behavior [14]. In addition, by investigating pilots’ research, Wang et al. showed that individual emotional intelligence can effectively improve safety behaviors [15]. Although previous studies help us to understand or predict the role of the human–machine interface, organizational factors, and individual factors in reducing human errors, there are still several problems with the research: first, most of the existing studies focus on pilots, drivers, and construction workers, and less consider mine workers under the smart mines system. As a safety-critical area, the safety of coal mining enterprises is of great significance. Second, humans as a member of the system cannot be studied unilaterally but should be investigated holistically from the perspective of interaction with specific scenarios.

The attitude–behavior process model, which states that attitudes can describe and predict behavior [16], has been widely used in the study of safety behavior. In addition, some related studies have shown that miners’ poor psychological states, such as poor attitudes and motivation, cause the highest frequency of human safety accidents [17,18,19,20]. Therefore, it is very important to study human-causal failures of coal miners based on the attitude–behavior model. Attitudes have two mechanisms to influence individual behavior [21,22]: in a normal state, individuals are in a position to think carefully and will weigh various options before making careful behavioral decisions, while in a busy or emergency state, individuals have no time for precise weighing, and their attitudes will immediately form a sense of perception of various situation and then react quickly to things. That is, situational awareness perception could have a mediating role between attitude and behavior. In addition, based on trait activation theory, the trait that an individual possesses in itself must be influenced by the environment that fits the trait to exhibit a certain behavior [23]. Zhang et al. argued that task complexity is important in the safety-critical area [24]. The complexity of the task as the core of human–machine interaction can, to a certain extent, increase one’s vigilance, thus awakening the trait of safety attitudes and making the individual more capable of maintaining continuous attention to detect changes in the surrounding environment.

Given this, considering the smart mine as a safety-critical area, its working environment is more complex and the work task has a certain degree of difficulty. Therefore, based on the attitude–behavior process model, this paper introduces two variables, situational awareness and task complexity, from the perspective of human–situation interaction and human–machine interaction and establishes a moderated mediation model to explore in-depth the mechanisms and boundary conditions of the role of safety attitudes in human errors, as shown in Figure 1. This study enriches the theory of safety attitudes and human errors, thus providing new explanations for how safety attitudes can reduce miners’ human errors.

2. Theory and Hypotheses

2.1. Safety Attitudes on Human Errors

Human errors are human actions or behaviors that exceed the accepted standards or permissible range required for the system to work, i.e., people do not perform according to the standards designed in the regulations when completing a task [25]. A smart mine is the IT platform for the statistical perception of production elements, processes, logistic support resources, etc., to realize the operation mode of automatic analysis and processing of information sharing and transmission. In a smart mine, “human” behavior is highly uncertain, and its reliability and stability are much lower than those of machine facilities because human behavior is constrained by environmental, physiological, and psychological factors [26,27]. Numerous studies have shown that human factors, especially human errors such as personnel violating safety procedures and conscious and unconscious misuse of operating rules and regulations, are the main causes of failures and problems in coal mine production [28,29,30].

Safety attitudes as an intrinsic psychological trait are a state of mental preparedness that individuals analyze and judge to avoid the occurrence of damage [31,32]. As a safety-critical area of smart mines, risk is everywhere, and individuals need to analyze and judge the safety status of their environment to guide them to take the right actions. Attitude triadic theory divides safety attitudes into three aspects: safety cognition, safety emotion, and safety behavior tendency [33]. Among them, safety cognition refers to the knowledge, concept, or idea that individuals form about the object through their perception, memory, thinking, and imagination [34]; safety emotion refers to the emotional reaction that individuals have to the object, such as optimism or negativity, importance or contempt [35]; safety behavior tendency refers to the state of readiness before individuals make a reaction in the face of the object [36]. Based on the triadic theory [37], we can find that safety attitude as an individual’s subjective safety evaluation affects people’s attention allocation and is the state of preparation for their behavior. Moreover, in terms of behavioral theory, Tan et al. argued that workers’ safety attitudes have a direct relationship with safety behaviors [38], and Meng et al. further suggested that increased attention to safety and workers’ safety attitudes plays a positive role in promoting safe citizenship behaviors [39]. Therefore, this study concludes that employees with good safety attitudes have time to think, are more accurate in their safety judgments, and are more able to comply with regulations, thus effectively reducing human error. Based on the above analysis, this paper proposes the following hypotheses.

Hypothesis 1 (H1).

Safety attitudes have a significant negative effect on human factors errors.

2.2. The Mediating Role of Situational Awareness

Situational awareness refers to the perception, understanding, and prediction of the meaning of each element in the environment in a specific space and time, i.e., some reactions and characteristics that individuals show internally in the face of the external changing environment, which is an important factor influencing people’s decision making in the changing information environment [40,41,42]. Thus, employees with high situational awareness are more able to perceive changes in the surrounding environment and thus make correct decision-guiding behaviors. Situational awareness of employees is easily influenced by the individual’s knowledge structure, cognitive ability, and physiological and psychological status [43]. Specifically, Monazzam noted that good attitudes of individuals enable them to notice changes in their environment and shift their attention from less important tasks to more critical matters [17]. For example, individuals in emergencies do not have more time to think, and a good safety attitude enables individuals to quickly perceive changes in the situation and pay attention to more detailed parts; on the contrary, when individuals have poor safety attitude, they tend to not take changes in their surrounding environment seriously and thus fail to effectively avoid mistakes. Thus, in either case, safety attitudes affect people’s attention allocation and change their perception of what is around them, thus affecting situational awareness.

Based on the cognitive psychology information processing model [40], Endsley divided situational awareness into three levels of observation, comprehension, and prediction: level one, perceiving the elements in the surrounding environment, selectively noticing and briefly remembering the key elements [43]; level two, judging and understanding the key elements briefly remembered in level one, and further comprehensively understanding the current situational consisting of each element [44,45]; level three, combining Levels 1 and 2 and existing empirical knowledge to make judgments about the development trend of the work environment, predict the future state, and make the best decision in the current situational [46]. Thus, it can be seen that situational awareness guides individuals to make appropriate behavioral choices to a certain extent and reduces individual human errors. In addition, Lu et al. [47] also emphasized that situational awareness provides the basis for behavioral decision-making. Therefore, situational awareness may reduce human-causal errors to some extent.

Based on the above analysis, this study suggests that the effect of safety attitudes on human errors may be mediated by situational awareness and safety attitudes reduce human errors by influencing individuals’ attention to reason about situation information, thereby helping to engage in rational behavior. As a result, the following hypothesis was formulated.

Hypothesis 2 (H2a).

Safety attitudes have a significant positive effect on situational awareness.

Hypothesis 2 (H2b).

Situational awareness has a significant negative effect on human errors.

Hypothesis 2 (H2c).

Situational awareness plays a mediating role between safety attitudes and human errors of employees.

2.3. The Moderating Role of Task Complexity

With the development of interaction psychology, Tett and Guterman [23] argued that individuals who possess a trait must first be influenced by an environment that is compatible with the trait to exhibit a behavior. Task environments can amplify or inhibit trait effects on performance, depending on whether the relevant task environment is trait-related. The complexity of the task precisely defines the process of human–machine interaction and generates unique contextual features that are important factors in shaping employees’ work processes and outcomes [48,49]. Therefore, this study considers task complexity as a boundary condition of the “individual trait–situational awareness” relationship.

However, there is some debate about the moderating role of task complexity. Starting from resource conservation theory, some scholars argued that task complexity makes individuals focus on their primary task to the exclusion of other information, which is detrimental to safety [50]. In contrast, some scholars, in exploring the effects of task complexity on employees in nuclear power plants, found that task complexity increased individual alertness, thus making employees more concerned about safety during work [24]. Similar to nuclear power plants, smart mines are safety-critical areas with inherently dangerous work scenarios and complex and variable work tasks. Therefore, the safety task is its main task. Miners are more alert when faced with complex tasks and thus allocate more of their attention to safety tasks [51]. Combining trait activation theory, this high vigilance mentality and attention allocation trigger the role of safety attitudes in promoting situational awareness. Thus, when task complexity is high, safety attitudes will play a greater role in promoting and maintaining situational awareness. Conversely, when task complexity is low, the role of safety attitudes may not be as prominent. Specifically, when task complexity is low because most members are not committed to improving safety, resulting in a low organizational safety climate, people who hold good safety attitudes, therefore, tend to hide their traits, which they may perceive as making the trait of holding safety attitudes unappreciated and unrecognized or even denied [52]. In addition, the low task complexity does not provide members with clear safety information, resulting in members not being more alert, which is not conducive to improving situational awareness and reducing human-caused errors. Based on this, the following hypothesis is proposed in this paper.

Hypothesis 3 (H3).

Task complexity plays a positive moderating role between safety attitudes and situational awareness; specifically, the stronger the task complexity, the stronger the positive effect of safety attitudes on situational awareness.

Based on the above logic and assumptions, a mediated model with moderation is proposed in this study. Specifically, based on the attitude–behavior model, when miners have good safety attitudes, they are more able to perceive changes in the surrounding scenario, thus reducing human errors. As the core of human–machine interaction, task complexity may moderate the mediating role of situational awareness. Therefore, we propose the following hypothesis.

Hypothesis 4 (H4).

Task complexity positively moderates the indirect effect of safety attitudes on employees’ human errors through situational awareness, i.e., the more complex the task, the stronger the effect of safety attitudes on employees’ human errors through situational awareness; conversely, the simpler the task, the weaker the effect.

3. Method

3.1. Data Collection and Sampling

A sample of full-time staff working under the smart mine system in China was collected for this study. Before the questionnaire was administered to the staff, they were informed that the data collected from the questionnaire they were to complete would be used for academic research only, that there were no correct or incorrect answers to these questions, and that their personal information would not be disclosed to ensure the anonymity and confidentiality of the questionnaire. To reduce common methodological biases associated with self-reported data, participants were asked to complete a survey twice at two different points in time. In the first questionnaire (Time 1), workers provided demographic information and answered questions about safety attitudes and task complexity. Approximately one month later (Time 2), they were asked to answer a second questionnaire, which included questions about situational awareness and human errors.

At Time 1, 438 questionnaires were distributed, and 320 valid questionnaires were collected after discarding invalid questionnaires (response rate = 73.1%). At Time 2, we redistributed new questionnaires to 320 participants. In this process, we used a unique code to match the two questionnaires of 320 participants, and finally, 246 participants’ questionnaires were successfully matched (response rate = 76.9%). The demographic characteristics of the respondents in this study are described in

Table 1. Sample Demographics.

Characteristics	Category	Percentage (%)
Gender	Male	97.9
	Female	2.1
Age	18–25 years old	4.1
	26–35 years old	26.4
	36–45 years old	53.3
	46–55 years old	13.0
	Over 56 years old	3.2
Educational Background	Less than high school	33.7
	High school	34.6
	College	17.5
	Master’s degree	14.2
Working years	0–5 years	3.3
	6–10 years	22.8
	11–15 years	33.4
	16–20 years	28.5
	More than 20 years	12.0

.

3.2. Measures

To ensure the reliability and validity of variable measurement, most scales in this study used relatively mature English scales. Most scales in this study used a five-point Likert scale, except for the statistical variables, and the values ranged from 1 (strongly disagree) to 5 (strongly agree).

(1)

Safety attitudes: A scale developed by Seaboch was adopted [53]. It consists of 13 questions, five of which are about safety cognitive attitudes, such as “I think safety accidents at work can be prevented”, three questions about safety affective attitudes, such as “I am willing to wear safety protection equipment for work”, and five questions about safety behavioral tendencies, such as “Before starting work, I tend to check equipment and facilities for safety hazards”.

(2)

Human errors: The 13-item instrument created by Shakerian [54] was employed to measure miners’ human errors, with higher scores indicating higher levels of human errors in a person’s work. These included items such as “Have you ever started doing something before bringing or preparing its necessary tool due to a mental and job engagement?”.

(3)

Situational awareness: The 10-item scale created by Sneddon et al. [45] was adopted to measure miners’ situational awareness. A higher score demonstrated a higher level of situational awareness. Example items are “I find it easy to keep track of everything that is going on around me”.

(4)

Task complexity: Task complexity was measured using the perceived task difficulty questionnaire developed by Robinson [55], which measures coal miners’ perceived task complexity in five dimensions: difficulty, stress, confidence, interest, and motivation, with five items, such as, “I think my job task is a bit difficult”.

(5)

Control variable: To avoid the impact of demographic variables on the research results, this study set the following variables as controls—age, gender, education level, and working years.

3.3. Data Analysis Procedures

In this study, we applied SPSS 22.0 (IBM, Armonk, NY, USA), Mplus 8.0, and Hayes SPSS macro program Process 3.3 to analyze the collected data. The analysis was performed according to the following methods. First, AVE and confirmatory factor analysis were performed using SPSS 22.0 and Mplus 8.0, respectively, to ensure the reliability and validity of the data and to determine the measurement model. Then, descriptive statistics and correlation analysis of each variable were initially checked by SPSS 22.0. Finally, multivariate stepwise linear regression was conducted to test the mediating role of situational awareness and the moderating role of task complexity; our moderated mediation model was tested by Process 3.3 using a bias-corrected percentile bootstrap method. A bootstrap sample of 5000 was drawn to obtain 95% confidence intervals (CIs). If the 95% confidence interval excludes 0, it means that the effect is statistically significant.

4. Results

4.1. Reliability and Validity Analysis

The convergent validity of each variable was analyzed, and the average external variance (AVE) and composite reliability (CR) were compared. As shown in

Table 2. Results of reliability analyses.

Variable	Cronbach’s α	AVE	CR
safety attitudes	0.951	0.522	0.934
Situational awareness	0.934	0.507	0.911
human errors	0.947	0.513	0.932
task complexity	0.864	0.541	0.855

, the AVEs for safety attitudes, task complexity, situational awareness, and human errors were 0.522, 0.541, 0.507, and 0.513, respectively; the CRs for each variable were 0.934, 0.855, 0.911, and 0.932, respectively. All variables reached the critical values of AVE > 0.5 and CR > 0.7. In addition, the Cronbach’s α of all four variables is greater than 0.7, indicating that the questionnaire has good reliability.

Confirmatory factor analysis using Mplus 8.0 for safety attitudes, task complexity, situational awareness, and human errors.

Table 3. Results of confirmatory factor analyses.

	χ2	$\mathbf{df}$	χ${}^2/\mathbf{df}$	RMSEA	SRMR	TLI	CFI
Hypothetical model	967.031	773	1.296	0.032	0.042	0.970	0.972
Three-factor model	2043.543	776	2.633	0.081	0.104	0.806	0.817
Two-factor model	2814.310	778	3.617	0.103	0.116	0.690	0.706
Single-factor mode	3293.888	779	4.23	0.115	0.127	0.617	0.636

Note: hypothetical model represents safety attitudes, task complexity, situational awareness, and human errors; three-factor represents safety attitudes + task complexity, situational awareness, human errors; two-factor represents safety attitudes + task complexity + situational awareness, human errors; single-factor represents safety attitudes + task complexity + situational awareness + human errors.

shows that the factor indices were significantly better than other alternative models and met the criteria for acceptability (χ² = 967.031, $df=$773, χ${}^2/df=1.296$, CFI = 0.972, TLI = 0.970, SRMR = 0.042, RMSEA = 0.032). The four variables in this study belong to different constructs and have good discriminant validity.

4.2. Descriptive Statistics and Correlations

The means, standard deviations, and correlation coefficients in this study are listed in

Table 4. Descriptive statistics and correlations (n = 246).

	Mean	SD	1	2	3	4	5	6	7
1. Gender	1.17	0.373
2. Age	2.85	0.817	−0.011
3. Education	2.58	0.965	0.049	−0.029
4. Working years	3.24	1.039	−0.018	0.052	−0.861 **
5. Safety attitudes	2.87	0.977	0.044	−0.021	0.078	−0.077
6. Situational awareness	2.85	0.932	0.078	0.017	0.121	−0.113	0.536 **
7. Human errors	2.82	0.984	−0.047	−0.041	0.079	0.117	−0.662 **	−0.513 **
8. Task complexity	2.80	0.897	0.082	0.045	0.074	−0.052	0.074	0.587 **	−0.248 **

Note: n = 246, ** p < 0.01.

. The study had the following findings: (1) safety attitudes were negatively correlated with human errors (r = −0.662, p < 0.01); (2) safety attitudes were positively correlated with situational awareness (r = 0.536, p < 0.01); situational awareness was negatively correlated with human errors (r = −0.513, p < 0.01); (3) task complexity was positively correlated with situational awareness (r = 0.587, p < 0.01).

4.3. Results of Hypothesis Testing

4.3.1. Testing Results of Main Effects

Model 4 examined the effect of safety attitudes on employee human errors. As shown in

Table 5. Mediating role of situational awareness and moderating effect of task complexity.

Variable	Situational Awareness			Human Errors
	Model 1	Model 2	Model 3	Model 4	Model 5	Model 6
Gender	0.130	0.028	0.024	−0.056	−0.028	−0.026
Age	0.036	0.005	0.013	−0.073	−0.046	−0.065
Education	0.056	0.004	−0.009	0.123	0.136	0.135
Working years	−0.021	−0.039	−0.047	0.164	0.167	0.159
Safety attitudes	0.504 ***	0.469 ***	0.484 ***	−0.663***		−0.547 ***
Situational awareness					−0.536 ***	−0.229 ***
Task complexity		0.568 ***	0.558 ***
Safety attitudes × Task complexity			0.164 ***
R2	0.298 ***	0.591 ***	0.615 ***	0.450 ***	0.273 ***	0.483 ***
F	20.329	57.526	54.251	39.213	18.010	37.181

Note: n = 246, *** p < 0.001.

, safety attitudes were negatively correlated with human errors (r = −0.663, p < 0.001), indicating that employees with higher safety attitudes were less likely to make human errors. Therefore, hypothesis H1 was supported.

4.3.2. Mediating Role of Situational Awareness

The results of Model 1 indicated that a safety attitude had a significant positive effect on situational awareness (r = 0.504, p < 0.001), and hypothesis H2a was supported by the data. Model 5 showed that situational awareness negatively influenced human errors (r = −0.536, p < 0.001), and hypothesis H2b was tested. Model 6 adds the mediating variable situational awareness to Model 4, and it was found that the regression coefficient of safety attitudes on human errors decreases (r from 0.663 to 0.547), while the effect of situational awareness on human errors remains significant (r = −0.229, p < 0.001), indicating that situational awareness plays a partially mediating role between safety attitudes and human errors, and hypothesis H2c is supported by the data.

To further verify the mediating role of employee situational awareness between safety attitudes and human errors, this study tested the indirect effect between safety attitudes and human errors based on Hayes’ (2013) suggestion. Bootstrapping results for 5000 parameters showed that the 95% confidence interval for the indirect effect of safety attitudes on human errors through situational awareness was [−0.188, −0.054], excluding 0, indicating the presence of a mediating effect of employee situational awareness, again supporting hypothesis H2c.

4.3.3. Moderating Effect

As shown in Table 5, Model 3 tested the moderating effect of task complexity on the relationship between safety attitudes and situational awareness. The corresponding results showed that the interaction term between safety attitudes and task complexity positively predicted situational awareness (r = 0.164, p < 0.001). Hypothesis H3 was validated, suggesting that task complexity enhances the positive effect of safety attitudes on situational awareness. The moderating effect of task complexity is shown in Figure 2, based on the test results of Model 3 presented in Table 5.

4.3.4. The Moderated Mediating Effect

The purpose of the path analysis was to examine the size and differences in the mediating effects of situational awareness when task complexity was at different levels (i.e., high and low). As can be seen from

Table 6. Examination of the moderated mediating effect (n = 246).

	Indirect Effect	Standard Error	95% Confidence Interval
Low Task Complexity(−SD)	−0.077	0.024	[−0.126, −0.034]
High Task Complexity (+SD)	−0.145	0.043	[−0.231, −0.063]
Difference	−0.068	0.026	[−0.126, −0.023]

Note: n = 246, low task complexity has a mean minus 1 standard deviation while high task complexity has a mean plus 1 standard deviation.

, when task complexity is low, the indirect effect of safety attitudes on human errors through situational awareness has a value of –0.077 with a standard error of 0.024 and a 95% confidence interval of [−0.126, −0.034], which does not contain 0, indicating a significant indirect effect. When task complexity was high, the indirect effect of safety attitudes on human errors through situational awareness had a value of −0.145, a standard error of 0.043, and a 95% confidence interval of [−0.231, −0.063], which did not contain 0, indicating a significant indirect effect. The value of the indirect effect difference between the high and low conditions for task complexity was −0.068, with a standard error of 0.026 and a 95% confidence interval of [−0.126, −0.023], excluding 0, indicating that the indirect effect difference value was significant, i.e., the mediating effect being moderated was present, supporting hypothesis H4.

5. Discussion

Reducing human error is not only a means to ensure workplace well-being, but also to achieve human sustainability goals. The findings show that safety attitudes can reduce human error; situational awareness partially mediates the effect of safety attitudes on human errors. This result is supported by the literature [11,14,16,17,18,19,20], which shows that safety attitudes can significantly reduce human error. In addition, task complexity positively moderated the relationship between safety attitudes and situational awareness and positively moderated the indirect effect of safety attitudes on human errors through situational awareness. Previous literature on task complexity remains controversial [50], and this result shows the positive moderating effect of task complexity through an empirical study, which is the same as Zhang’s findings [24].

As per research performed in this study, all hypotheses proved to be relevant, and the proposed methodology is consistent with similar studies [15]. Safety attitudes are effective in increasing situational awareness, thereby reducing human errors and ensuring safety and sustainability in the workplace.

5.1. Theoretical Implication

Our study provides several theoretical implications for the field of human errors. First, although previous studies have shown that good safety attitudes can positively influence miners’ safety behaviors, no pathway studies have been conducted [56]. Based on the attitude–behavior model, this study analyzed the different paths of safety attitudes on human errors to fill this gap. Moreover, the analysis of different pathways broadens the understanding of safety attitudes and again verifies the important role of safety attitudes on safety behaviors.

Furthermore, this study explored the mediating mechanisms of situational awareness. With the advancement and development of smart mines, the work of miners has changed from manual operation in the past to human–machine interaction, and the cognitive characteristics of operational tasks have increased. Amid such changes, the miners’ ability to quickly process changes in environmental information becomes critical [2]. Therefore, this paper introduces the variable of situational awareness to provide insight into how individual safety attitudes in emergencies can reduce human errors through the transmission of situational awareness. On the one hand, it has important theoretical value for uncovering the “black box” between safety attitudes and human errors; on the other hand, the validation of the mediating role of situational awareness enriches the theoretical study of situational awareness and promotes the influence of the process of human–situation interaction on behavior.

Finally, this study discusses the boundary conditions of task complexity and provides new insights into how miners’ safety attitudes trigger their situational awareness and consequently reduce their human errors. Based on trait motivation theory, this study reveals the moderating role of task complexity from a human–machine interaction perspective [15]. The study shows that when task complexity is high, safety attitudes are more likely to trigger their situational awareness to reduce human errors. In addition, the findings deepen the knowledge of task complexity and push the human–machine interaction operation to a new level, and also enrich the research of trait activation theory in safety behavior, which opens up a new path for effectively reducing the occurrence of human errors.

5.2. Practical Implications

It has been widely recognized by academics that companies place great emphasis on the positive impact of employee safety attitudes. However, not all employees have a high level of safety attitudes, nor does holding a safety attitude prevent accidents from occurring. Therefore, it is important to find ways or means to stimulate employees’ safety attitudes so that they can maintain their focus on their daily work.

First, the coal mining industry is inherently a safety-critical and high-risk area compared to other industries. Therefore, organizational managers should select employees who are competent and able to maintain a good safety attitude toward their work. Second, performance standards and pay structures can be adjusted daily to complement good safety attitudes. This is more conducive to shaping employee safety attitudes and creating a good safety climate to avoid human errors occurring under routine conditions. Finally, it is necessary to emphasize the importance of safety attitudes and safety awareness in a crisis environment. Studies have shown that when task complexity is high, individuals are more alert and more able to awaken safety attitudes and thus perceive changes in the environment [57,58]. Therefore, the organization’s managers in their daily work should increase the publicity and education of high-risk industries in smart mines, to build the crisis awareness of employees, so that they can take the initiative to realize that they are in high-risk work so that employees can maintain a good state to reduce human errors.

5.3. Limitations of the Current Study and Avenues for Future Research

This study has several limitations. First, the mediation of situational awareness is one theoretical perspective to explain the effect of safety attitudes on reducing human errors, but there are other possible pathways for safety attitudes to influence human errors. In this regard, it is recommended that future studies further discuss other potential impact pathways. Second, this study reveals how task complexity evokes safety attitudes to reduce human errors for employees, but there is still some room for improvement. Specifically, it is recommended that future research delves into other variables that can evoke safety attitudes in employees; this can be carried out from the employee’s perspective as well as the external environment, such as leadership [59], or by considering multiple factors such as the organization, superiors, and subordinates simultaneously to explore how to effectively trigger safety attitudes. Finally, how to reduce human errors is another research direction for the future. The research method of this paper is the form of self-report through questionnaires, and the research results may not be objective. In the future, on this basis, the experimental research of scenario simulation will be discussed in depth.

6. Conclusions

Based on the attitude–behavior model, this study explored the effects of safety attitudes on human errors and investigated the partially mediating role of situational awareness and the moderating role of task complexity. The results of the study showed that individual safety attitudes can effectively reduce human errors through the partially mediating role of situational awareness, where task complexity plays a moderating role in the relationship between safety attitudes and situational awareness, with higher task complexity awakening human safety attitudes and thus generating situational awareness to reduce human errors. Our study contributes to the workplace safety literature by introducing trait activation theory, advancing our understanding of how safety attitudes are triggered in high-risk work environments to improve situational awareness to reduce human errors. From a practitioner’s perspective, our findings suggest that, in addition to ensuring task complexity, companies can also conduct safety attitude tests when recruiting the right employees for safety-critical positions, as people with good safety attitudes are better able to observe changes in scenarios to reduce human errors.