Promoting Sustainable Safety Work Environments: Factors Affecting Korean Workers’ Recognition of Their Right to Refuse Dangerous Work

Abstract

(1) Background: The right to refuse dangerous work (RTRDW) is essential for preventing industrial accidents and protecting worker safety in Korea. However, its use remains limited in practice. This study seeks to identify the factors hindering its activation across industries such as construction, manufacturing, and services, offering a comprehensive analysis beyond previous research. (2) Methods: A survey was conducted across key industries to assess five factors—safety behavior, communication, management commitment, education and training, and education and training—using structural equation modeling (SEM) to evaluate their influence on the exercise of RTRDW. (3) Results: The SEM model showed a good fit (χ² = 1151.333, p < 0.001, TLI = 0.978, CFI = 0.984, RMSEA = 0.05). The most significant factors influencing RTRDW were safety performance behavior and communication, while ambiguous regulations, poor training, and fear of job loss discouraged its use. (4) Conclusions: To improve RTRDW activation, clearer regulations, enhanced safety education and training, stronger management commitment, and better communication are necessary. Addressing these issues can help workers confidently exercise their right to refuse dangerous work, enhancing overall workplace safety. (5) Benefits: This study provides practical strategies for policymakers and industry leaders to promote safety, empowering workers to use RTRDW effectively and contributing to a safer work environment.

1. Introduction

While the number of fatalities from industrial accidents in Korea has shown a long-term declining trend, the industrial accident fatality rate in 2023 remains close to or above 1, with 0.98 deaths per 10,000 workers covered by industrial accident insurance [1]. This situation indicates that workers’ lives and health are still exposed to significant risks. A review of the characteristics of industrial accidents and fatal incidents in Korea reveals that many occur in small businesses with fewer than 50 employees [2]. Considering that small-scale businesses tend to conceal industrial accidents or are reluctant to report them, the actual occurrence rate of industrial accidents could be higher [3]. This accident rate has been influenced by the indirect-employment trend promoted by companies [4]. Although official statistics are unavailable, it is generally recognized that workers employed through outsourcing and performing subcontracted tasks have a higher rate of industrial accidents compared to directly employed, full-time workers of primary contractors [5]. Ensuring the safety of indirectly employed workers engaged in outsourced tasks requires, above all, the responsibility of the primary contractor and the realization of subcontractors’ rights [6].

To address these challenges, effectively prevent industrial accidents, and protect workers in the workplace, the Occupational Safety and Health Act (OSHA) was enacted [7]. Although this law primarily mandates the preventive duties of management, it also stipulates the right of workers to refuse work as a critical element of their rights concerning safety and health [8]. In Korea, the right to refuse work is stipulated in Article 52 of the Occupational Safety and Health Act, which grants workers the authority to refuse work and evacuate in the event of an imminent risk of an industrial accident. Additionally, if there is a reasonable basis to believe that such imminent danger exists, management is prohibited from dismissing or otherwise disadvantaging workers who have refused work and evacuated for this reason [9].

This right to refuse work serves as a critical measure to prevent accidents or injuries and protect workers’ lives and health by enabling them to immediately halt work when their safety is compromised. It contributes to the overall reduction in industrial accidents. However, in practice, few instances are reported where workers have exercised this right, contrary to the law’s intent [10]. According to the Ministry of Employment and Labor’s “Survey on the Actual Status of Safety and Health Management Regulations at Workplaces”, the actual usage rate of the right to refuse work by workers within the past year is only 16.3 % [11]. Furthermore, the right to refuse work is often utilized as a post-incident measure following a severe accident rather than as a preventive tool to avoid accidents [12]. This reality indicates that working conditions are not adequately guaranteed in practice, and the protection of workers’ rights remains insufficient [13]. Therefore, the effective implementation of the right to refuse work can be a key factor not only in safeguarding workers’ lives and health [14] but also in ultimately promoting sustainable economic growth [15].

While the right to refuse dangerous work (RTRDW) is a crucial element in protecting workers’ life and health, there are challenges in ensuring and exercising this right effectively. Therefore, it is essential to explore ways to activate RTRDW. Previous studies have analyzed the conditions for exercising RTRDW or examined the reasons for its underutilization. Building on these studies, we conducted a quantitative statistical analysis of the relationships between the factors that can enhance RTRDW. Understanding these relationships provides valuable insights for implementing structural strategies to activate RTRDW across industries in Korea, offering a broader approach compared to previous studies that focused on specific sectors or cases.

2. Literature Reviews

Although the right to refuse dangerous work (RTRDW) is designed to protect workers’ life and health, there are significant challenges in realizing this right. Various studies have highlighted these issues and proposed solutions, but these efforts are often fragmented. This calls for a more integrated and quantitative approach. In the literature review section, we examine studies related to RTRDW, analyzing the factors that can enhance its activation. By reviewing previous research, we explore the gaps that this study seeks to address in order to better understand the conditions for successfully implementing RTRDW.

Kim (2015) defined who needs protection under occupational safety and health laws and explored what is required to ensure such safety. The Occupational Safety and Health Act protects workers, clearly defining workers and emphasizing that responsibility lies with the employer. Although workers’ right to refuse work for safety reasons is legally guaranteed, it is not fully utilized in practice to prevent accidents. Kim attributed this to the requirement that “imminent danger of industrial accidents” must be met to exercise the right, but without clear criteria, workers risk civil or criminal liability if it is determined they misjudged the situation. Kim stressed that clear standards are needed to secure this right [4].

Kwon (2023) presented two perspectives on determining “imminent danger of industrial accidents”: one based on the employer’s duty to ensure work refusals when required (objective view) and the other based on the worker’s judgment (subjective view) [13]. Through a recent case [16] law analysis, Kwon argued that the worker should be the decision-maker regarding “imminent danger”. However, due to the lack of codified standards for this right, legal disputes are likely to continue [17].

Park (2015) argued through legal interpretation that the RTRDW is not an individual worker’s right but a collective one, where workers can refuse work and evacuate if they believe, based on reasonable grounds, that there is an imminent danger. This collective action, according to Park, contributes to the protection of workers’ lives and safety [18].

Jeong (2014) claimed that the RTRDW could lead to labor disputes, as there are no clear legal grounds or requirements to enforce this right [19].

Cho (2014) noted that when workers exercise their right to refuse work, they must present reasonable grounds that the employer can trust. This burden of proof on the worker can lead to additional legal liabilities if the employer does not acknowledge it. Cho also emphasized that although employers are required to take action to eliminate dangers when workers exercise their right to refuse work, in reality, employers often pressure workers to resume work without addressing the hazards, citing productivity concerns. Cho stressed the importance of employer commitment to safety to ensure the effective realization of workers’ rights [10].

Cho (2013), through interviews with safety managers and labor unions, found that the misuse of the RTRDW sometimes undermines its original intent of preventing accidents. Cho argued that a specific organization is needed to assess whether the right has been appropriately used and called for clear regulations to distinguish between work refusals and strikes [8].

Shim et al. (2023) conducted a survey on the implementation of employers’ responsibility to refuse work during imminent danger, such as typhoons, and identified ways to activate the right to refuse work. The survey confirmed that employers were using the right in workplaces, and Shim emphasized the need for regulatory oversight, guidance, and promotional safety education to ensure the active use of this right. Additionally, Shim called for clear regulation and procedures for implementing work refusals [20].

Song (2022) conducted a survey in the construction industry to analyze workers’ awareness of the RTRDW and identified issues through accident case analysis. The survey revealed that 81% of respondents had never exercised their right to refuse work, with many workers either unaware of or indifferent to the right. Many companies did not include information on the right in their safety education and training, and the lack of clear regulations for the right raised concerns about potential misuse. Analyzing three fatal accident cases, Song pointed out worker-unsafe behavior and a lack of communication between workers and supervisors as reasons for the failure to exercise this right [21].

Coulson et al. (2014) investigated how communication barriers between miners, supervisors, health and safety officers, and labor unions in South Africa affect the effective exercise of the right to refuse work through interviews and surveys. The survey analysis was conducted utilizing SPSS software, version 15. They found that while workers were aware of their right, supervisors or safety officers were more likely to exercise it than workers themselves. Workers who had better communication with supervisors exercised the right more frequently, highlighting the importance of communication in ensuring workers’ safety rights [22].

Coulson et al. (2019) also conducted interviews and discussions to understand South African miners’ awareness of the right to refuse work and the barriers they face in applying this right to ensure safety. The study revealed that while workers generally understood the concept of the right, they were less familiar with formal procedures, and fear of retaliation, as well as a productivity-driven culture, discouraged them from exercising it. Particularly, the emphasis on productivity was a major obstacle to the exercise of the right, underscoring the need for a balanced safety culture [23].

As shown in

Table 1. Comparison with previous Studies. ○ represents the factor discussed in previous studies.

Authors (Year)	Factors
	Management Commitment	Communication	Education and Training	Regulation and Procedure	Worker Safety Behavior
[4]				○
[13]				○
[18]				○
[19]				○
[10]	○
[8]				○
[20]			○	○
[21]		○	○		○
[22]		○
[23]	○		○

, the existing literature identifies key factors influencing the RTRDW, such as management commitment, communication, safety regulations and procedures, safe worker behavior, and safety education and training. However, previous studies have only addressed some of these factors, and few have analyzed their structural relationships in an integrated manner.

Based on the existing literature, the following hypotheses were established:

Hypothesis H1.

Management commitment positively influences the awareness of the right to refuse dangerous work (RTRDW).

Hypothesis H2.

Communication positively influences the awareness of RTRDW.

Hypothesis H3.

Safety regulations and procedures positively influence the awareness of RTRDW.

Hypothesis H4.

Safety education and training positively influences the awareness of RTRDW.

Hypothesis H5.

Safety behavior positively influences the awareness of RTRDW.

To analyze the structural relationships of the proposed factors, we closely examined the safety culture that fosters workers’ awareness and actions regarding their rights, as argued by Coulson et al. (2019) [23]. Safety culture is a key concept in the literature on accident prevention and analysis [24], and it has been defined by various researchers. The concept was first introduced after the Chernobyl disaster and was described as “the essence of safety culture is the means by which both organizations and individuals pay close attention to safety” [25]. Studies on safety culture typically focus on defining the concept [26] or measuring its maturity model [27]. The factors related to the right to refuse work are similar to those used to assess the maturity of safety culture.

To create a safe working environment, it is essential to improve safety culture. To improve safety culture, it must first be measured; however, since direct measurement is challenging, safety climate—an observable and measurable reflection of the current state—can be used as a proxy [28,29]. While a safety climate reflects the situation at a specific moment, a safety culture represents a more enduring phenomenon [30], and safety climate scores are often used to indicate the maturity of safety culture [31]. The five dimensions used maturity models in safety culture are management commitment, communication, education and training, regulations and procedures, and employee involvement [32,33,34,35,36]. Examining the relationships among these five dimensions, Moon [37] found that management commitment, communication, safety regulations and procedures, and safety education and training, which represent safety climate, all influence safety behavior. Tappura et al. [38], in a survey of workers in the chemical industry, demonstrated that management commitment positively affects employee safety behavior through safety education and training, with safety education and training directly impacting worker behavior. Therefore, this study aims to explore the internal relationships among these five dimensions and how they influence workers’ awareness of the RTRDW.

3. Methods

3.1. Sample and Data Collection

This study focuses on industries in Korea, specifically targeting the construction, manufacturing, and service sectors, which are representative of the broader Korean industrial landscape. The data were extracted from the “Industrial Safety and Health Survey”, which assesses the safety and health conditions of workplaces in Korea and provides essential data for policy formulation. The survey is conducted every three years and covers approximately 7000 workplaces. For this study, the most recent data from the 2021 survey were used [39].

A total of 7268 respondents participated in the survey, and the distribution of respondents was categorized by industry sector. Workplace size was determined by total project value for the construction industry and by the number of regular employees for manufacturing and service industries. Among the respondents, 3255 were from manufacturing, 2511 from services, and 1502 from construction. Most of workplaces had fewer than 300 employees or project values under 300 million KRW, representing smaller-scale workplaces. This sample reflects the reality of small businesses in Korea, where exercising workplace rights is often challenging due to subcontracting relationships (

Table 2. Analysis of demographic information about respondents.

Industrial	Size (Person, Millions KRW)	N (%)
Manufacturing	Total	3255
	20–49	2277 (70)
	50–99	608 (18.7)
	100–299	264 (8.1)
	300–999	96 (2.9)
	Above 1000	11 (0.3)
Service	Total	2511
	20–49	1477 (57.9)
	50–99	559 (21.9)
	100–299	358 (14.0)
	300–999	140 (5.5)
	Above 1000	16 (0.6)
Construction	Total	1502
	50–120	717 (47.8)
	120–300	350 (23.3)
	300–500	122 (8.1)
	500–1000	139 (9.3)
	Above 1000	174 (11.6)

).

3.2. Measurement

Through the literature review, five factors influencing workers’ awareness of the right to refuse dangerous work (RTRDW) were identified. These factors were linked to the data collected from the “Industrial Safety and Health Survey” and used in the analysis. In this context, the latent variables represent the factors affecting workers’ awareness of the RTRDW, while the observed variables are expressed through survey questions (

Table 3. Latent and observed variables.

Latent Variables	Observed Variables	Questionnaire
Management Commitment	Q1	The management demonstrates a strong commitment to ensuring workers’ health and safety.
	Q2	The management consistently prioritizes safety over other considerations.
	Q3	Safety is regarded as a critical value by the management.
Communication	Q4	Sufficient opportunities are provided for employees to discuss safety-related issues during meetings.
	Q5	Open dialog about safety concerns is actively encouraged within the workplace.
	Q6	Employee feedback on health and safety matters is regularly sought.
Education and Training	Q9	Ample opportunities are provided for employees to participate in comprehensive health and safety training sessions.
	Q10	Health and safety training is highly effective in preventing workplace accidents.
Regulation and Procedure	Q7	Systematic and well-structured safety regulations and procedures are established.
	Q8	Safety regulations and procedures are proven to be effective in reducing workplace accidents.
Safety Behavior	Q11	Employees consistently follow safety procedures and work standards.
	Q12	Employees proactively engage in improving workplace safety without needing direct instruction.
RTRDW	Q13	Employees have the right to refuse work if they deem it to be dangerous.

). All observed variables were measured using a 5-point Likert scale. The research model, developed based on previous studies, is shown in Figure 1.

3.3. Data Analysis Techniques

Data analysis was conducted using SPSS Statistics 27 and Amos 18.0. Descriptive statistical analysis was employed to summarize the collected data and examine the distribution and characteristics of key variables, utilizing frequency analysis, mean, and standard deviation. Subsequently, factor analysis was performed to identify the factors influencing workers’ awareness of the RTRDW, enabling the evaluation of items constituting the latent variables. To ensure the internal consistency and reliability of each identified factor, Cronbach’s alpha coefficient was applied. Correlation analysis was also conducted to explore the relationships between workers’ awareness of the RTRDW and potential influencing factors. For a more rigorous examination of the relationships between variables, structural equation modeling (SEM) was employed [40], which assessed both the direct and indirect effects of each factor on workers’ awareness and evaluated the overall fit of the theoretical model.

The analysis identified the key factors influencing workers’ awareness of the right to refuse work and explored the internal relationships among these factors. This served as the foundation for proposing strategies to enhance the effective implementation of this right.

SEM, a model that integrates measurement and structural components, was utilized to accurately investigate the relationships between variables. Based on the study’s objectives and theoretical framework, the factors influencing workers’ awareness of the right to refuse work were defined as latent variables. Hypotheses were formulated to examine how these latent variables influence awareness. Observed variables, drawn from survey or questionnaire data, were used to measure the latent variables. A confirmatory factor analysis (CFA) was conducted to validate how well the observed variables represented their respective latent constructs.

Model fit was evaluated using a range of indices, including a chi-square (χ²) test, Tucker–Lewis Index (TLI), Comparative Fit Index (CFI), and Root Mean Square Error of Approximation (RMSEA). A model was considered to fit well if the chi-square test p-value was below 0.05, TLI exceeded 0.9, CFI exceeded 0.9, and RMSEA was below 0.08 [41]. In cases where the model did not demonstrate adequate fit, paths were adjusted or variables were added or removed to improve the model specification [42]. The fit of the structural model was assessed using the same criteria as for the measurement model.

Once the structural model demonstrated satisfactory fit, a path analysis was conducted to examine the magnitude and direction of the relationships between key variables. Path coefficients were estimated to determine the strength and directionality of the relationships, illustrating the extent to which independent variables influence dependent variables [43], thereby allowing for an assessment of the relative importance of each factor. An SEM is a robust analytical tool that enables the estimation of measurement errors, assessment of the reliability and validity of measurement instruments, evaluation of model fit, and simultaneous measurement and prediction [44].

4. Results

4.1. Measurement Models Using EFA and CFA

The measurement model is primarily conducted using Confirmatory Factor Analysis (CFA), a method of factor analysis used when the relationships between latent and observed variables are established through theory or previous research. However, since the latent variables used in this study lack pre-existing scales, an Exploratory Factor Analysis (EFA) was first conducted to develop new measurement instruments, followed by a CFA to validate the measurement model.

Various mathematical procedures can be employed to perform factor analysis, but the general objective is to estimate the relationships between variables and factors [45]. Common estimation methods include Maximum Likelihood Estimation (MLE) and the Least Squares Method [46]. In this study, factors were extracted using the Least Squares Method, and a Varyimax rotation technique was applied to maximize the variance of the variables for interpretability.

To assess the suitability of the factor analysis, a Kaiser–Meyer–Olkin (KMO) test was conducted, yielding a value of 0.937. A KMO value greater than 0.5 indicates that the data are suitable for factor analysis [47]. Additionally, Bartlett’s test of sphericity confirmed the appropriateness of the factor analysis model, with a significance level of 0.000, indicating that the data were suitable for factor analysis [48]. Throughout the factor analysis process, iterative validation was performed on all subsets of conceptual items or questions in the survey. In the initial analysis phase, subsets of items within questions were considered, and through repeated analysis, variables with no strong correlation with other items were intended to be eliminated [46]; however, no items were ultimately omitted.

The factor analysis based on the measured variables resulted in the identification of five latent variables. The CFA was conducted to include an appropriate number of factors, aiming to maintain a simple structure while adequately explaining the data variance [49]. All variables demonstrated a communality of 0.7 or higher, and the factor loadings and variances for each variable are presented in

Table 4. Factor analysis results.

Variables	Factor 1	Factor 2	Factor 3	Factor 4	Factor 5	Communality
Q1	0.781					0.783
Q2	0.830					0.877
Q3	0.843					0.848
Q4		0.758				0.799
Q5		0.827				0.859
Q6		0.656				0.734
Q9			0.798			0.916
Q10			0.606			0.802
Q7				0.792		0.865
Q8				0.764		0.852
Q11					0.758	0.828
Q12					0.816	0.868

. Factors were grouped based on the absolute value of factor loadings being 0.4 or higher, and five factors were ultimately extracted, with all factor loadings having an absolute value of 0.4 or above.

4.2. Mean, SD, and Correlations Among Variables in This Study

The model in this study consists of six key latent variables: management commitment, communication, education and training, regulations and procedures, safety behavior, and workers’ refusal to work. All observed variables were measured using a 5-point Likert scale. To estimate the Structural Equation Model (SEM), the data must follow a normal distribution; therefore, the mean, standard deviation, skewness, and kurtosis were examined to verify this assumption.

According to the criteria for normal distribution proposed by West, Finch, and Curran (1995), all variables exhibited skewness below 2 and kurtosis below 7, confirming that the data follow a normal distribution [43] (

Table 5. Normality test results.

Variables	Mean	Standard Deviation	Skewness	Kurtosis
Q1	4.54	0.665	−1.541	2.878
Q2	4.56	0.660	−1.607	3.156
Q3	4.58	0.647	−1.630	3.214
Q4	4.35	0.740	−0.964	0.696
Q5	4.31	0.745	−0.886	0.535
Q6	4.22	0.809	−0.840	0.399
Q9	4.26	0.784	−0.809	0.224
Q10	4.23	0.786	−0.775	0.248
Q7	4.20	0.820	−0.798	0.190
Q8	4.24	0.781	−0.789	0.240
Q11	4.25	0.766	−0.822	0.531
Q12	4.14	0.827	−0.826	0.597
Q13	4.29	0.769	−0.924	0.642

).

A reliability and correlation analysis were conducted among the key variables. To assess reliability, internal consistency was examined using Cronbach’s α [50]. The results indicated that all variables had Cronbach’s α values above 0.7, confirming a satisfactory level of reliability.

Additionally, correlation analysis is useful not only for hypothesis testing in exploratory studies but also for providing an overview of the relationships between variables. The analysis revealed that all correlations between latent variables were significant at the p < 0.001 level (

Table 6. Reliability analysis and correlation matrix.

Variables	Cronbach’s α	Correlation Coefficient
		MC	CM	ET	RP	SB
MC	0.901
CM	0.863	0RE.671 **
ET	0.814	0.605 **	0.701 **
RP	0.851	0.582 **	0.665 **	0.752 **
SB	0.817	0.594 **	0.623 **	0.666 **	0.688 **
RTRDW	-	0.512 **	0.575 **	0.584 **	0.576 **	0.661 **

** p < 0.01, MC: management commitment; CM: communication; ET: education and training; RP: regulation and procedure, SB: safety behavior; RTRDW: right to refuse dangerous work.

).

4.3. Meserement Model Verification

The fit indices of the final measurement model were satisfactory, with χ² = 1072.066 (p < 0.001), TLI = 0.976, CFI = 0.984, and RMSEA = 0.057. The CFA results indicated that the paths from the latent variables—management commitment, communication, education and training, regulations and procedures, and safety behavior—to the observed variables were all significant at the 0.001 level, as shown in

Table 7. Confirmatory Factor Analysis results.

Latent Variables	Observed Variables	Estimate		Standard Error	p-Value
		$B$	$\mathit{\beta }$
Management Commitment	Q1	0.946	0.812	0.011	p < 0.001
	Q2	1.058	0.914	0.01	p < 0.001
	Q3	1	0.882
Communication	Q4	0.936	0.826	0.012	p < 0.001
	Q5	0.96	0.841	0.012	p < 0.001
	Q6	1	0.807
Education and Training	Q9	0.943	0.805	0.012	p < 0.001
	Q10	1	0.852
Regulation and Procedure	Q7	1.033	0.853	0.012	p < 0.001
	Q8	1	0.867
Safety Behavior	Q11	1	0.849
	Q12	1.037	0.816	0.014	p < 0.001

. Here, Estimation B represents the degree to which each variable influences the regression model, while β represents the standardized values of these effects.

4.4. Structual Model Verification

The final Structural Equation Model is presented in Figure 2. The model fit indices were satisfactory, with χ² = 1151.333 (p < 0.001), TLI = 0.978, CFI = 0.984, and RMSEA = 0.051. As shown in

Table 8. Path analysis results.

Path	Estimate		Standard Error	p-Value
	$B$	$\mathit{\beta }$
Management → Safety Behavior	0.215	0.189	0.016	p < 0.001
Management → Education	0.196	0.177	0.017	p < 0.001
Communication → Management	0.655	0.748	0.011	p < 0.001
Communication → RTRDW	0.216	0.183	0.019	p < 0.001
Communication → Education	0.677	0.701	0.17	p < 0.001
Regulation → Safety Behavior	0.363	0.392	0.0.35	p < 0.001
Education → Regulation	1.018	0.915	0.014	p < 0.001
Education → Safety Behavior	0.348	0.338	0.042	p < 0.001
Safety Behavior → RTRDW	0.705	0.594	0.020	p < 0.001

Management: management commitment; Education: education and training; Regulation: regulation and procedure.

, the results of the structural model analysis indicate that both communication and safety behavior have a direct and positive influence on workers’ awareness of their right to refuse dangerous work (β = 0.183, β = 0.594, respectively). These results support Hypotheses H2 and H5, confirming their direct effects. However, while the other hypotheses did not demonstrate direct effects, it was found that they exert indirect influence on the activation of the right to refuse dangerous work.

Management commitment has a positive and significant impact on education and training (β = 0.177) and safety behavior (β = 0.189). This suggests that higher levels of management commitment are associated with improved education and training and enhanced safety behaviors among workers within the workplace.

Moreover, higher levels of communication were found to enhance management commitment (β = 0.748) and positively influence education and training (β = 0.701). Additionally, as training and education are strengthened, there is a corresponding increase in the levels of safety regulations and procedures (β = 0.915) and safety behavior (β = 0.338). The analysis further demonstrates that well-established regulations and procedures lead to higher levels of safety behavior (β = 0.392).

5. Discussion

This study employed Structural Equation Modeling (SEM), providing a more robust validation than traditional regression analysis typically used in survey research, thus enhancing the reliability of the results. The findings underscore the necessity of emphasizing workers’ right to refuse dangerous work (RTRDW) in order to create a sustainable working environment and protect workers’ health and safety. Although this right is legally guaranteed in Korea, it is rarely exercised in practice. Obstacles to exercising the RTRDW include the absence of clear regulations supporting such actions, which causes workers to fear retaliation, and insufficient education about the right, leaving many unaware of its existence. Furthermore, a lack of management commitment to safety often results in pressures prioritizing productivity over safety. Additionally, poor communication between supervisors and workers may hinder the actualization of workers’ right to refuse dangerous work, even when exercised by workers. This study statistically analyzed the factors that act as barriers to workers exercising this right.

The two most critical factors directly influencing workers’ awareness of the RTRDW are communication and safety behavior. This finding suggests that both the government and employers should prioritize improving workplace communication and enhancing workers’ safety behaviors.

The findings of Song (2022) support this study’s conclusion that communication between supervisors and workers positively impacts awareness of the RTRDW (Communication → RTRDW, β = 0.183). Actual case studies, where supervisors ignored urgent danger reported by workers, leading to accidents, highlight the importance of communication in fostering awareness of the RTRDW.

Workers who voluntarily comply with regulations and procedures, actively participate in improving workplace safety, and demonstrate proactive safety attitudes and behaviors are more likely to engage in preventive maintenance activities [51]. Moreover, workers who are aware of safety issues and take proactive safety actions are more motivated to report and raise concerns about safety issues [52]. By promoting safety behavior, safety culture is enhanced, which in turn increases the likelihood of securing workers’ rights.

Safety behavior, which can have a significant impact on accidents, refers to a set of actions taken by individuals to ensure safety, such as following regulations and procedures or participating in safety activities to prevent risks associated with safety performance [53]. For workers to consistently follow regulations and procedures and exhibit safe behaviors, management must assign safety responsibilities, provide education and training, and implement rewards to foster a strong safety culture within the workplace [54]. This aligns with the finding that factors influencing workers’ awareness of the RTRDW are closely related to those that impact safety culture.

Kim et al. (2021) found that construction safety education and training can enhance safe behaviors. Based on the current study’s findings (Education → Safety Behavior, β = 0.338), incorporating information on the RTRDW into construction safety education and training is expected to raise both safe behavior and workers’ awareness of this right [55].

Hong et al. (2016) pointed out that regulatory content and its applicability can influence workers’ violations. Safety standards that are too general or outdated may not be relevant to current tasks. Additionally, there are cases where regulations exist but are difficult to comply with or are not used effectively [56]. This aligns with previous research, which suggests that while the RTRDW is legally guaranteed, the lack of clear criteria makes it difficult to exercise. Therefore, clear and actionable regulations are necessary for workers to realize this right (Regulation → Safety Behavior, β = 0.392). Moreover, workers may violate regulations because they do not understand their importance. This highlights the need for education and training to explain the purpose and principles behind safety regulations and procedures (Education → Regulation, β = 0.915), ensuring compliance through understanding rather than simple instruction [57,58].

Barling and Zacharatos (1999) identified leadership as a critical organizational determinant of safety [59]. Leaders play a key role in creating a safe environment and motivating employees to work safely (Management → Safety Behavior, β = 0.189) [60,61]. Management commitment can improve not only workers’ safety behaviors but also the educational environment. Choi et al. (2016) also demonstrated that management’s interest positively influenced training and education [62]. When management actively participates in safety education and training, it can improve the quality of instructors, the content, and the training environment. Such involvement also enhances participation and engagement from both workers and supervisors (Management → Education, β = 0.177).

6. Conclusions

This study was conducted to address the issue that, despite the potential for workers’ right to refuse dangerous work (RTRDW) to play a crucial role in preventing industrial accidents and protecting workers’ safety and health in Korea, it has not been fully utilized in practice. The findings indicate that the main reasons for the under activation of RTRDW include ambiguous regulations, insufficient safety and health training, lack of management commitment, poor communication, and negative perceptions among workers, such as fear of job loss.

Previous studies have criticized the vague criteria for exercising RTRDW and explored factors that hinder its activation in specific industries. However, this study makes a significant contribution by systematically identifying the factors influencing RTRDW based on a survey of key industries in Korea, including construction, manufacturing, and services. Additionally, this research integrates the factors previously explored in the literature and utilizes statistical methods to propose a structural strategy for enhancing RTRDW.

Using Structural Equation Modeling (SEM), the final model demonstrated satisfactory fit indices, with χ² = 1151.333 (p < 0.001), TLI = 0.978, CFI = 0.984, and RMSEA = 0.05. The analysis revealed that, among the five factors examined, safety performance behavior and communication significantly influence the activation of RTRDW.

The results of this study align with findings from various other studies. However, rather than drawing isolated conclusions, an integrated structural strategy is needed. Based on the final model, the following structural strategies are proposed:

Management Commitment and Support: Management’s commitment to safety should encourage workers’ safety behaviors. Management must foster a safety-conscious environment and provide support for employees to perform their tasks safely.

Improving Education and Training: Management should enhance the quality and environment of safety education and training to sustain workers’ awareness of safety behaviors. Interviews with field experts during this study revealed that involving management in safety education and training is one of the most effective ways to improve the quality of instruction, content, and the environment. Additionally, if management mandates the inclusion of the RTRDW in safety education and training, instructors will be more likely to address the risks that workers perceive in the workplace.

Establish Clear Regulations: For workers to recognize and exercise their right to refuse work, clear regulations are necessary. Although the RTRDW is legally guaranteed, the lack of specific and relevant standards hinders its application. Therefore, regulations should be improved to ensure that workers understand and can exercise their rights, supported by training that clarifies the purpose and principles behind these regulations.

Enhancing Communication: Communication between different levels of the organization should be strengthened. In some cases, workers reported imminent danger but were ignored by supervisors, leading to accidents. This demonstrates the critical role of communication in addressing safety issues. Improved communication can also enhance the content and quality of education and training, leading to improved safety behaviors and greater awareness and exercise of the right to refuse dangerous work.

Fostering a Safety Culture: Workers who voluntarily comply with safety procedures and actively participate in improving workplace safety are more likely to engage in preventive maintenance activities and report safety concerns. Thus, improving the safety culture is essential to enhance workers’ safety performance behaviors.

6.1. Theoretical Implications

This study contributes to the literature on occupational safety by identifying key factors that influence the activation of the RTRDW, such as management commitment, clear safety regulations, education and training, and communication. It provides a theoretical framework that enhances the understanding of worker safety behaviors and safety culture, while also calling for further research in diverse national and industrial contexts.

6.2. Practical Implications

For practitioners and policymakers, this research emphasizes the importance of establishing clear safety protocols, demonstrating strong safety leadership, providing continuous safety education and training, and fostering open communication. These actions are essential to empower workers to exercise their RTRDW, prevent industrial accidents, and create a sustainable safety culture in the workplace.

6.3. Limitations and Future Research

This study was conducted primarily in the context of workplaces in South Korea, which may limit the generalizability of the findings to other countries and industries. Differences in legal, cultural, and industrial factors across countries may affect the exercise of workers’ RTRDW, which were not fully captured in this research. Therefore, future studies should investigate the factors influencing the RTRDW in various countries and industries. Such research would allow for more generalizable conclusions that account for cross-national and cross-industry differences. Additionally, further research is needed to explore how emerging technologies, such as smart technologies and AI, impact the exercise of the RTRDW. Understanding how technological advancements influence workers’ safety perceptions and their ability to assert their rights could enhance the practical relevance and applicability of future studies.