Although individuals are responsible for developing and maintaining behaviors that reduce safety and health risks, individual behaviors are simultaneously influenced by factors at external levels. Some experts argue that an ecological approach is better suited for at-risk populations [16], such as mineworkers whose environment increases their vulnerability to certain injuries and diseases [17]. In the case of mine safety training and skill maintenance, the SEM is an informative framework because mineworkers first learn and then apply competencies in different environments (i.e. the training facility and actual mine site, respectively). The SEM therefore allows for a focused analysis of how these various environments might influence the mastery and maintenance of critical skills. Each level of the SEM is discussed below.

The intrapersonal or individual level includes characteristics that influence behavior, such as knowledge, attitudes, skills, and beliefs [18]. Current mine safety training takes an intrapersonal level approach in which the individual is the target for developing competencies [14,19]. The interpersonal level provides role definitions and personal relationships, such as contact with family, friends or coworkers, which may influence behavior [19]. For example, because mineworkers often work together in crews on a consistent basis, whether peers encourage or discourage safety behaviors could have a significant influence on behavior. The organizational level also can facilitate and support individuals' willingness to change behaviors [14]. The organization itself can be a target for many health and safety initiatives, including local rules and policies to ensure employees' safety and health. An example includes adopting worksite practices that support preventative care, such as a smoking cessation program to assist in the prevention of respiratory problems. The community level includes social norms and values that exist among collective groups that can impact structures and the behaviors within those structures, such as the propensity to take risks and willingness make safer decisions [18,19]. Strategies at this level are typically designed to impact the processes and proximal rules within a given work system. Examples specific to the mining community include nonverbal communication signals with cap lamps or task training for a particular machine. Last, the societal level includes cultural context and regulatory policies that facilitate healthier behaviors [19]. State mining agencies and the MSHA are societal-level factors that influence worksite policies and ultimately the work behaviors of mine site personnel. For instance, the required training we discuss throughout this paper is one of the regulatory factors within the mining industry.

Some researchers argue that, although theoreticians often express interest in and use the SEM, practitioners rarely take advantage of this model's utility [14,15,20]. Using the SEM to consider potential shortcomings in self-escape training processes is an applicable and novel approach within the mining industry. The five-level framework allows for an examination of various factors within an individual's environment, including peer networks and organizational rules [21]. In addition, because a combination of both individual-level and environmental resources are needed to promote the mastery and maintenance of self-escape KSAOs, it is particularly important to understand where problems exist within the contextual levels of an ecological system in order to offer appropriate recommendations [15].

The SEM has been applied to help develop, implement, and evaluate training at the community level in an effort to promote health and safety behaviors within specific populations. Examples include smoking cessation programs, safer-sex practices, and addressing community violence from multiple levels [14–19]. We considered SEM as a potentially valuable evaluative tool in the case of mine safety training as the SEM has been efficacious in revealing barriers to knowledge transfer and behavior change in other health and safety issues. Interviews with mine safety trainers were therefore analyzed in an effort to reveal potential weaknesses in relation to training content, skill development and assessment, and skill maintenance within and among the five ecological levels.

The research questions posed in this study emerged from two key issues: (1) mastering and maintaining self-escape KSAOs during and after training; and (2) consideration of whether an ecological framework can be used to identify limiting factors across various levels of training systems. With this perspective in mind, this study sought to utilize the SEM as a lens through which to analyze interviews with mine safety trainers to reveal factors that hinder mineworkers' mastery and maintenance of self-escape KSAOs.

A semi-structured interview guide was developed that asked participants to respond to questions related to: (1) methods for teaching and assessing self-escape competencies; (2) best methods and the feasibility of these methods; (3) self-escape competencies on which mineworkers need better training, assessment and remediation; and (4) obstacles to stronger assessment methods. The interview questions were developed from a review of materials from mine safety training regulations and the literature, and input from mine safety training SMEs [1]. Participants were asked the same set of questions, but the interviews were flexible in that participants could discuss the specific mine safety issues that they perceived to be the most critical in terms of training competence and emergency preparedness.

The semi-structured interview guide was reviewed and approved prior to data collection [23]. Existing contacts were utilized to recruit participants, employing a convenience sampling method [24]. Potential participants were contacted either by telephone or e-mail. No one who was contacted declined to participate. Once the agreement to participate was secured, two researchers traveled to the participant's place of employment to conduct the interview. Informed consent was obtained prior to each interview. Participants were assured that their responses would be confidential and that their responses were voluntary.

The participants were nine mine safety trainers, considered to be SMEs in the area of mine safety training, and they had a combined training experience of 154 years. The participants resided in four states and had experience conducting new miner training and annual refresher training. Five participants were trainers for specific training institutions, one for a safety department within a mine, and three for state or federal organizations. Personal training experience ranged from 4 years to 40 years (mean experience, 17.1 years; standard deviation, 10.8).

Two researchers were present for each interview. One researcher conducted the interview and took hand-written notes while the other researcher observed and took hand-written notes. Each interview lasted 2–3 hours. After the third interview, responses started to become repetitive and little new information was presented, indicating early saturation of content [25]. Due to the specificity of the content and the homogeneity of the sample, this was not unexpected. For this reason, after all of the participants who were initially recruited to participate completed their interviews, recruitment ended.

Interview notes were typed by the note-taker immediately following each interview to form a loose transcript for each participant. Transcripts were cross-referenced with the interviewers' notes for content accuracy. In order to assess how the ecological model could be used to reveal gaps and inform ideas for improving the self-escape portion of mine safety training, each interview transcript was analyzed and coded. Three researchers worked concurrently during the coding process to develop a codebook that consisted of themes, codes, categories, and examples. Prior to the coding process the researchers met to discuss data analysis steps and coding rules to ensure consistency of interview coding. Upon drafting the data analysis framework, researchers read through and coded the interviews independently. Researchers continued to meet to confirm that the coding process was going smoothly and to discuss any new themes or codes that were emerging. Four data analysis meetings were held prior to when the codebook was finalized. This iterative process helped to ensure reliability of coding across each interview.

General rules and procedures for qualitative analysis from a theoretical framework were used as guidance [26,27] throughout the following steps. (1) Familiarization of the data: Researchers became acquainted with the raw data by reading and re-reading the transcripts to identify key ideas and recurring topics. (2) Identification and application of a theoretical framework: Researchers used an ecological framework to identify the key issues and concepts by which the data could be analyzed and referenced. The text was coded according to SEM levels: individual, interpersonal, organizational, community, and societal. Identifying the five ecological levels within each interview resulted in data that were in “manageable chunks for subsequent retrieval and exploration” [27]. (3) The assignment of initial codes: After researchers identified and agreed on the assignment of SEM levels, they noted patterns within the data chunks that were subsumed within the SEM levels. Researchers wrote notes in the margins of the interview transcripts to begin identifying factors that trainers referenced within each ecological level. The researchers regularly met to discuss the patterns that were emerging. (4) Focused codes and organization: The researchers reread their handwritten notes and assigned codes and categories based on the repetition of text within each interview. Quotes from each interview were collated to help support saturation of the pattern and eventually conceptualize a theme to summarize the factors that were present across the levels of the SEM. Once the researchers agreed upon the patterns and codes that were emerging throughout the interviews, these themes became part of the codebook. Refer to Table 1 for an excerpt from the codebook. (5) Interpretation: Researchers used the final codebook to look for connections between themes and illustrate factors in safety training that need attention to improve self-escape training systems and processes.

Results indicate that effective communication and collaboration may be lacking within the mining system. These results are debriefed within each SEM level.

The second theme that emerged from the ecological analysis is the lack of leadership at each level of the SEM. Each level and its respective weaknesses are discussed below.

A third limitation that emerged from an ecological analysis of the data was the amount of responsibility and personal accountability that people are willing to take on in order to ensure that the mining workforce is prepared to self-escape.

First, the communication and collaboration factors present at each level of the SEM indicate that, rather than placing blame on the individual, trainer, or industry, the collaborative relationships within, among, and between all of these groups are necessary to improve workforce competency and performance. If more communication occurs between mine site leadership and safety trainers, between safety trainers and mineworkers, between mine site leadership and mineworkers, etc., a more accurate view of the knowledge and skill gaps of our nation's coal mineworkers may emerge. Specifically, enhanced communication between leaders of mine organizations and safety trainers may identify gaps in learning and stimulate site-specific training for skills that are more difficult for mineworkers to master. For example, during discussions between safety trainers and mine site leadership, deficiencies in map reading and familiarity with alternative escapeways may surface as a problem. These skills can be addressed in more detail during training, and organizations can specifically discuss alternative escapeways during weekly safety meetings or create a quarterly drill in which mineworkers are forced to use an alternative escapeway.

Trainers were not confident that work crews were proactively communicating with each other after training sessions ended. After starting work at a specific mine, work crews should communicate to establish common knowledge of escape aids. For instance, Trainer 2 said: “All escape decisions should be made prior to a disaster. You work with these people every day so you need to have that discussion.” Increased communication and teamwork may help to prevent the deterioration of skills over time. The leadership at mine sites may need to organize such activities to allow work crews to practice decision-making scenarios together.

The results from this study indicate that in order for safety training to fulfill their intended purpose to develop self-escape KSAOs, leaders must be present at all levels of the mining system. If mineworkers are encouraged to be positive leaders and are provided with leadership opportunities during and after safety training, they can build leadership skills and model safer behaviors for their co-workers. The respective leadership roles at the mine organizations, however, are critical in terms of refining, mastering, and sustaining mineworkers' KSAOs.

There are too many skills to be effectively taught during training and so they must be acquired on the job at a specific mine. Participation from organizations during and after training may facilitate this follow-up process. For instance, when talking about evacuating during an emergency, one trainer said: “I do not typically teach miners how to decide the best evacuation routes. This is handled by someone from the specific mine” (Trainer 1). Similarly, as the circumstances change at a particular mine, onsite training must occur for mineworkers to maintain mine-specific self-escape KSAOs. As one trainer noted, students must have the ability to escape from primary and secondary escapeways as the mine develops and changes. These results indicate that it is crucial for mine site leadership to take an active role to ensure that skills become second nature through extra training on the job. In one example of this, in Australian coal mines the underground foremen are charged with supervising employees and the mine operations within a designated area of the underground mine. These individuals receive more comprehensive training in self-escape in order to take the lead in evacuating crews of miners [28]. A leadership model that at least targets certain job positions and/or tasks may be a positive start to further develop leadership skills at United States mines. Lastly, trainers said that part of being a leader at a mine site is establishing a culture that holds workers more accountable for unsafe behaviors and empowers workers to make safer choices. Trainers expressed an overall sense that mine safety culture is improving, which leads to the final theme that emerged in the data—responsibility and accountability.

The results across ecological levels showed a gap in the responsibility and accountability that individuals possessed to ensure mastery of the KSAOs needed to self-escape. A process of iterative teaching and learning is needed to build a more competent workforce. First, results indicate that mineworkers need to be accountable for their own self-escape competence. Trainers indicated that mineworkers tend to be reliant on their peers both during and likely after safety training. When an emergency occurs, therefore, some individuals may not know how to escape.

Trainers and organizations also need to hold themselves accountable for providing accurate and usable information to the mining workforce in order to develop and maintain those KSAOs. Organizations should provide ongoing training opportunities for employees to help maintain skills and teach specific nuances of the mine in which they work. In addition, trainers should maintain their own competencies in the mine environment and stay up-to-date with new mining technology (e.g., refuge alternatives, communications, etc.) so they can discuss such information during safety training.