Click for next page ( 64


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 63
4 Decision Making P icture a situation where a coal miner is working and encounters smoke and flying dust. What does the miner do? Don the self-contained self- rescuer (SCSR)? Radio a supervisor? Find out where the smoke is coming from? Immediately set out for a place of safety? Which step comes first? In this chapter, we explore the role that decision making plays in response to a potential mine emergency. To effectively respond to a mine emergency, miners must have the psychological tools to detect signs that an emergency exists and then use these tools to make effective decisions about how to act. In short, effective decision making is critical for ensuring that miners can extract themselves to a place of safety in an emergency. We take a human-systems integration approach to understand decision making in a mine emergency. Our intent is to highlight knowledge about human strengths and limitations in the context of an interactive system of people, equipment, and their environment that will be useful for preparing miners for self-escape in the event of a mine emergency (see Henricksen et al., 2008). We focus on the miner, giving an overview of psychology and neuroscience work documenting what happens in the brain and body in stressful situations. We then use this knowledge to elucidate the factors that cause people to make decisions (good or bad) that could influence self-escape. There are several different approaches to the investigation of decision making. For example, the normative approach describes how decisions ought to be made—the optimal, rational decision given a fully informed decision maker. In contrast, the descriptive approach characterizes how 63

OCR for page 63
64 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES people actually make decisions, the biases they bring to the table, and the different factors that influence the decisions they make (Shafir and Tversky, 2002). Finally, naturalistic decision making can be loosely thought of as an extension of the descriptive approach. It addresses how people make decisions in demanding environments (e.g., uncertain and changing envi- ronments, stressful situations, or time pressure; Klein et al., 1993). Natu- ralistic decision making also accounts for how decision-making practices may change as a function of a person’s experience, their work culture, etc. In this chapter, we take a more descriptive than normative approach. Our goal is to highlight the importance of considering decision making and decision science research more broadly for enhancing self-escape in a mine emergency. Certain themes recur throughout this chapter. One such theme is the importance of the use of decision science research (arising from the fields of psychology and neuroscience) to inform and shape how miners deal with emergency situations. Miners need to be knowledgeable of typical warning signals and how to efficiently and accurately determine if a true emergency exists. Decision science research can shed light on the types of information miners might miss and the common mistakes that may be made in emergency situations. To the extent typical emergency scenarios can be predicted, decision science research can also inform the development of emergency protocols and training procedures so that mine workers are able to make effective decisions and take an appropriate course of action to escape to a place of safety in the case of an emergency. A second theme in this chapter is communication. In mining emergen- cies, miners not only need to be able to communicate with one another underground, but they also need to send and receive information with communication centers on the surface. Effective communication involves an understanding of the cognitive capacities of individual miners and how information is conveyed from the surface personnel to miners underground and back and among those miners who are underground. Because technol- ogy is a key asset in these situations, successful communication also involves an adequate understanding of how miners use technology to communicate and the limitations of that technology. Communication is also driven by the emphasis placed on receiving timely and adequate information at the organization level. Often there are brief opportunities to intervene and stop a potential emergency event from building to the next level. These brief opportunities are referred to as “golden minutes” (see, e.g., Horne et al., 1995). The adequate exchange of information and fluent communication is necessary to take advantage of these “golden minutes” opportunities.

OCR for page 63
DECISION MAKING 65 DETECTING A MINING EMERGENCY Risk is an inherent component of underground coal mines. Therefore, miners must draw a distinction between routine hazards and those that require self-escape. How is such a distinction drawn? Sensitivity and Bias One model to conceptualize the detection of a mining emergency, drawn from a rich literature in psychology of attention, is called signal de- tection theory (Green and Swets, 1966). Signal detection theory is driven by the general premise that almost all decisions people make take place in the presence of some uncertainty. Signal detection theory provides a language for representing decision making in the presence of uncertainty. As such, it may be useful for thinking about the decisions miners make (and the fac- tors that influence those decisions) when faced with information that there may be an emergency. Consider a situation in which a miner must decide whether there is an emergency situation. There are four possible outcomes (see Figure 4-1). The miner’s goal is to accumulate information that will increase the likelihood of getting either a hit or a correct rejection, while reducing the likelihood of an outcome in the two error boxes. Signal detection theory can be used to conceptualize people’s ability to detect an actual emergency. There are two important factors: A miner’s sensitivity (i.e., the ability to detect an actual emergency) and bias (i.e., the predisposition to say whether there is an emergency or not). Sensitivity and bias are inde- pendent and thus can be influenced by separate factors. In biomedical fields, sensitivity (as in the sensitivity of a test for a par- ticular illness) relates to a probability of the test revealing a positive result Actual Emergency Yes No Yes Hit False Miner’s Alarm Decision Correct No Miss Rejection FIGURE 4-1  Possible outcomes faced by a miner. Figure 4-1

OCR for page 63
66 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES given that a patient is ill. Making an analogy to mine emergencies, sensitiv- ity refers to the likelihood of detecting that there is a mine emergency when there is in fact one occurring. To increase sensitivity, miners must become better at perceiving indica- tors of emergency situations. With proper training, a miner is more likely to learn what cues indicate a real emergency situation and which do not. This means that with more training, miners will be able to acquire more (and more reliable) information (Willingham, 2001). Sensitivity is also impacted by situational awareness. Situational awareness is, in simplest terms, know- ing what is around you. Situational awareness can be defined as involving (a) the perception of a situation’s elements in time and space, (b) compre- hension of their meaning, and (c) projection of a near future status for the condition in question (Endsley, 1988; Endsley and Garland, 2000). Related to the idea of situational awareness is the concept of informa- tion uncertainty (i.e., when a person does not have all the information pertaining to the situation at hand). Recognizing when information un- certainty exists—and the steps that need to be taken to obtain necessary information—is imperative for the effective diagnosing of a potential emer- gency situation. Bias is influenced by characteristics of the miner, the self-escape task, and the equipment/technology in place to aid self-escape. As an example, people have a tendency to hold an optimism bias in which they initially ignore signs that there is a problem (Sharot et al., 2007). This bias could impact the immediacy with which miners recognize a problem exists and diagnose it as an emergency. In addition, if miners do not trust the safety equipment (e.g., SCSRs, carbon monoxide [CO] monitors) or have adverse expectations for what it will be like to use the equipment, they may be biased not to acknowledge that there is an emergency situation at hand. Bias is also impacted by organizational and external factors. For example, if there are external penalties for false alarms (e.g., lost productivity that could adversely impact the mining company or even peer influence with other miners not putting on their SCSRs), this may make miners less likely to respond, a conservative bias. Bias is also influenced by whether there is the presence or absence of an organizational safety culture, with the latter implicitly creating pressure not to false alarm (see Chapter 5). In sum, many factors can influence bias and sensitivity. The factors outlined above are meant to provide an illustration of system and miner characteristics that can influence how miners respond to indicators of a potential mine emergency. Next, we consider two specific examples of an emergency situation where the signal detection theory framework can be used to better un- derstand a miner’s decision-making process. In the first example, a miner encounters smoke and dust (and unbeknownst to the miner, toxic levels of

OCR for page 63
DECISION MAKING 67 CO in the atmosphere). The miner must decide whether or not the envi- ronmental factors encountered mean one should don the SCSR. Correctly diagnosing the atmosphere as unbreathable would be a “hit” (see Figure 4-1). In contrast, determining that the air is still breathable would be a “miss.” Importantly, the miner’s decision is not only influenced by sensitiv- ity to cues about air quality but also by a bias not to acknowledge that there may be an emergency situation. If the miner does not have faith that the SCSR will work, is not properly trained on the equipment, or is fearful of negative consequences for using an SCSR when it might not be absolutely necessary, he or she might be biased not to acknowledge the gravity of the atmospheric conditions and thus conclude that the SCSR does not yet need to be used—which in this situation would be a “miss.” In other words, even though a miner might be highly sensitive to environmental cues that the atmosphere is dangerous, the bias not to acknowledge the gravity of the situation may push one to conclude that there is no need to don the SCSR when in fact it would be the correct response. In the second example, a CO monitor goes off and the miner must de- cide if one should self-escape. The CO monitor was actually triggered from several pieces of diesel equipment operating nearby and thus self-escape is not necessary. If the miner decides that there is no actual emergency situa- tion because information is received from fellow crew members about the diesel equipment, then this would be a “correct rejection.” Note, however, that even without knowledge of the diesel equipment the miner might be biased to assume that everything is fine. This could be because the CO monitor has false alarmed several times in the past. In this second example, a miner’s bias would lead to the correct decision—pushing the miner to cor- rectly reject the CO monitor alarm as a sign of an emergency. This second example also illustrates the concept of “alarm fatigue,” a situation in which people learn to ignore alarms or possible environmental signs that there might be a problem because they routinely occur and usu- ally do not indicate an imminent threat to safety. Although, in the above example, alarm fatigue did not lead to a failure to respond to an emergency, there are other situations where it could cause a miner to ignore important signs that a problem has occurred. It is imperative to make miners aware of the possibility of alarm fatigue and create training conditions that provide miners with knowledge about the mine and their equipment and technology so that they can successfully determine which alarms and abnormal envi- ronmental conditions (e.g., smoke) are most likely to represent an imminent threat to safety. Such training can provide information about circumstances under which miners should err on the side of caution (i.e., have a bias to say an emergency exists) because the benefits of a “hit”—correctly diagnosing a problem—far outweigh the negative consequences of a false alarm. One such example is donning an SCSR when there is smoke. Because breath-

OCR for page 63
68 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES ing in this environment can be potentially very harmful, miners should be trained with a bias to don an SCSR when environmental conditions dictate it is likely needed (and not risk waiting to put it on until it is too late). In summary, the above examples illustrate that it is not just a miners’ ability to interpret cues in the environment regarding whether self-escape is necessary, but also one’s bias for being willing to say that an actual emer- gency exists or not. Being aware that both sensitivity and bias can influence decision making for self-escape is important for devising the best training methods to prepare miners for possible emergency situations. Expertise One way to train emergency detection is to help miners see their envi- ronment through the eyes of a seasoned miner or “expert miner.” Identify- ing the specific characteristics that constitute an expert miner is a difficult task. However, for the purposes of the present discussion, expert miners might be viewed as those nearing the end of a lifelong mining career (as opposed to those individuals who have only recently entered the work force). The expertise view (see Klein et al., 2013) is designed to allow less experienced miners to discover what an expert miner would identify as an emergency situation and why. Research on expertise demonstrates that highly knowledgeable individuals tend to classify situations based on their underlying causes while novices tend to be side-tracked by more trivial features. The bottom line is that a novice or inexperienced miner (or even a seasoned miner who has not been properly trained) may be missing im- portant information needed to classify an event as an emergency or not. There are techniques that can be helpful in eliciting expert knowledge (e.g., how a seasoned miner might act or make decisions in a potential emergency situation), such as verbal protocol analysis (Ericsson and Simon, 1999). Verbal protocol analysis is intended to capture the information an expert attends to when generating a decision or course of action rather than a description or explanation of what they are doing—the latter which may change by instructions to think aloud. Verbal protocol analysis is designed to simply help externalize the thoughts experts might otherwise keep inter- nal. As such, it can be a useful method for ascertaining the implicit wisdom of expert miners. It should be noted that experts do not always perform better than their novice counterparts. When an expert’s goal is to predict the mistakes a nov- ice may make, they often do this less well than novices themselves (Hinds, 1999). This is because experts often have trouble introspecting on their own performance knowledge (Beilock, 2010). It is also the case that when situations are ill-structured where a situation is not familiar, and it is hard to predict what will happen given initial problem cues, experts often do no

OCR for page 63
DECISION MAKING 69 better than their novice counterparts in interpreting these cues (Devine and Kozlowski, 1995). However, in situations where the information a miner encounters means there is a high probability of a certain event (e.g., smoke indicating a fire) and, given this information, there is an easily recogniz- able course of action (e.g., fighting the fire, deciding to self-escape), experts tend to outperform less experienced individuals in terms of diagnosing a particular situation and taking the appropriate course of action. In these situations, verbal protocol analysis may be advantageous in capturing this expert knowledge. Two classic studies on the psychology of expertise demonstrate that experts tend to classify situations (especially well-structured situations) based on underlying causes while novices are often side-tracked by trivial information that can lead them down the wrong solution path. In the first study, physics professors (experts) and undergraduate physics students (novices) were asked to sort a number of physics problems based on the characteristics they deemed most important (Chi et al., 1981). Novices tended to categorize the problems by surface features of problems whereas experts classify according to the major underlying physics principle govern- ing problem solution. In a coal mining situation, this could manifest itself as an inexperienced miner concluding that, when a CO alarm goes off, this means that there is a fire (and, in turn, if the novice miner finds there is no fire, and then ignores the correct alarm). In contrast, a more experienced miner would understand that a CO alarm could be triggered from a variety of underlying sources (e.g., a faulty alarm, a fire, nearby diesel equipment, or an accurate reading of gases from some other source). The ability to recognize that there are multiple underlying causes of the same alarm is important for determining what other information one needs to obtain to make the best decision about how to react. In the second study, Lesgold et al. (1988) assessed expertise in di- agnosing X-rays. First- and second-year medical residents (novices) and radiologists (experts) viewed a series of X-ray pictures and verbalized their diagnoses. The expert radiologists quickly evoked a schema (a mental model) for the probable diagnosis. They then brought in additional infor- mation to test their diagnoses (to try and both confirm and disconfirm—see discussion of confirmation bias errors below). Critically, they changed or altered their diagnosis as more details were discovered. In contrast, the medical residents (novices) did not apply the appropriate or complete confirmation tests to the problem schema they invoked. Furthermore, the residents’ schema was usually based on surface features of the X-ray and did not change easily with new or contradictory information. Two important qualities of expert performance can be taken from the above-mentioned work and can be incorporated into training for self-escape

OCR for page 63
70 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES in mine emergencies. The goal is to help miners to classify a situation ap- propriately and act in the most successful way to facilitate self-escape. • Expert performance is based on an extensive knowledge base and the organization of this knowledge in such a way that experts are able to recognize important underlying themes in a problem. This entails that experts see meaningful patterns of information where novices do not (Chi et al., 1981). • Experts have strong self-monitoring skills and metacognitive abili- ties, especially in well-structured situations. Experts are more ac- curate at judging the difficulty of the problems they encounter and noticing where their thinking might have gone awry. This allows them to flexibly update their mental model of the situation when new or contradictory information is encountered (Lesgold et al., 1988; Kruger and Dunning, 1999). AFTER A MINING EMERGENCY IS DETECTED Noticing a potential oddity in the environment is largely subserved by an area of the brain called the prefrontal cortex. The very front part of the human brain that sits above our eyes—the prefrontal cortex—is the seed of thinking and reasoning abilities (Beilock, 2010). Once there is a realization that something is amiss, a variety of brain and body reactions occur in re- sponse to a potentially stressful situation. For instance, adrenalin increases in the bloodstream which results in several physiological responses such as a racing heart, sweaty palms, and muscles preparing for action. Cortisol is also secreted, which helps keep the heart racing and blood sugar up. Registering that there is an emergency can also lead to worries about the situation and its consequences. These worries can overwhelm a per- son’s working memory, which governs one’s ability to think clearly in the moment, take in new and important information, and to make reasoned decisions (Wang et al., 2005; Beilock, 2008). Working memory is defined as a transient memory store involved in the control of a limited amount of information immediately relevant to the task at hand (Miyake and Shah, 1999). In simpler terms, working memory can be thought of as a flexible mental scratchpad that allows people to work with whatever information is inside consciousness. Working memory also helps people attend to some in- formation while ignoring other information (Baddeley, 1986; Engle, 2002). When working memory is compromised in stressful situations, decision making can be impacted. The concept of working memory in the context of the larger framework on human information processing is shown in Figure 4-2, a very general construal of human information processing with information bombarding

OCR for page 63
DECISION MAKING 71 Retrieving information Perception External Attention Working Long-Term World Memory Memory Storing Examples include information - general knowledge - specific episodes - techniques Control from long-term memory (e.g., experts might attend to different perceptual information than novices) FIGURE 4-2  General view of human information processing. a person from the outside world. At any given moment, people attend to Figure 4-2 some of what is around us and ignore other information. The informa- tion that is attended to enters working memory. Here, working memory is charged with the task of making sense of this new information in the con- text of what is already known (i.e., stored in long-term memory). As such, working memory plays an important role in the decision-making process. It represents a person’s ability to work with whatever information is held in consciousness, match it to past experiences, and generate an appropriate course of action. It follows then that if working memory is compromised, a person may perform at a less-than-optimal level (e.g., make poor decisions or select an inappropriate course of action). Research demonstrates that simply making people aware of common internal reactions in stressful situations (e.g., sweaty palms, beating heart) can make these reactions less distracting (Jameison et al., 2010). It’s also the case that training people to view their stress response as a sign of challenge rather than doom can lessen the negative impact of physiological arousal on effective thinking and reasoning (Mattaralla-Micke et al., 2011). One reason for this effect is that normalizing these responses makes people less likely to dwell on them. Dwelling on them further limits the working memory needed to be effective decision makers in stressful situations. In addition to the stress signals generated from the body, it is important to note that CO poisoning—which is a danger in some underground coal mines—can impact brain functioning. Specifically, CO poisoning is thought to cause difficulty in making decisions and processing information, key functions of the working memory system (Cohen, 2012). In nonstressful situations, working memory works in concert with emotional processing. However, when working memory is compromised, people’s decisions can be unduly influenced by emotional processes, which can lead to poor outcomes. This occurs because there are, generally, two

OCR for page 63
72 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES ways that people make decisions. One way relies heavily on mental re- sources, such as working memory. It is more systematic and analytic. The other way is based more on affective and emotional processes (Sloman, 1996; Stanovich and West, 2000; Kahneman, 2003). When people are in stressful situations, worries tend to co-opt working memory, leaving only the more affective processes to govern decision making. This condition may result in decisions that put other miners or oneself in danger, such as going back for friend, even though there are very clear indicators that this could put the rescuer and possibly other miners in extreme danger. Training min- ers to be aware that these emotion-driven decisions occur, and when they are most likely to occur, can provide them with better tools to understand their behavior and make optimal decisions in stressful mining emergencies. This science may inform training. For example, a training exercise could be developed in which miners encounter fictitious situations in which they might have an impulse to go back into a mine to rescue others. If the consequences (both positive and negative) of such a decision are made clear, the miner may be better able to make the most appropriate decision in the moment. Such education could also address cultural norms that dictate that miners must stick together, regardless of the consequences. Miners need to be taught that cultural norms may push them to make decisions that are inherently risky and driven by emotions. Miners need to be trained to consider all of the possible options in these situations. Note, a compromised working memory is not the only source of poor decision making. As an example, having an inappropriate procedure for donning an SCSR represents an error in long-term memory that could lead to problems with self-escape. The miner, here, is not compromised because of reduced working memory resources, but because he or she learned the wrong steps to begin with. Or, a miner may make the wrong decision about how to act based on analogy to a past circumstance that was similar in terms of surface features (e.g., a CO alarm) but not in underlying cause (e.g., a fire versus a source-unspecified gas leak). A lack of knowledge may lead to a particular course of errors (e.g., the miner finds there is no fire, so ignores the alarm, even though it is correctly diagnosing air problems), that could be avoided with training that provides a more detailed knowledge base of common environmental signs of problems, their underlying causes, and sensible courses of action. We turn to the issue of knowledge acquisition in more detail below and to the development of optimal training practices in Chapter 6. DECISION MAKING FOR SELF-ESCAPE As noted throughout this chapter, effective decision making is a criti- cal component of successful self-escape in a mine emergency. Importantly,

OCR for page 63
DECISION MAKING 73 effective decision making is not simply based on in-the-moment choices, but is also based on the long-term accumulation of knowledge and skills. Knowledge of Equipment and Technology Miners need to have extensive experience with the use of breathing ap- paratus, such as the SCSR, and they need to be able to use this equipment in conditions that are not optimal (including, but not limited to, poor vis- ibility). Miners also need to be able to effectively operate these devices in stressful environments that compromise the working memory one would otherwise have at his or her disposal. One way this can be accomplished is by training miners so that the use of these devices is automatic or habitual. It is believed that skill acquisition progresses through distinct phases characterized by differences in the memory operations supporting perfor- mance (Beilock and Carr, 2001). In the early stages of learning, skill ex- ecution is supported by working memory and monitored in a step-by-step fashion (Fitts and Posner, 1967; Anderson, 1993; Proctor and Dutta, 1995). However, procedural knowledge specific to the task develops with practice. Procedural knowledge operates largely outside of working memory and does not require constant control (Anderson, 1993; Beilock et al., 2002). Thus, in contrast to earlier stages of performance, once a skill becomes relatively well learned, attention may not be needed for the step-by-step control of execution. One can think of procedural memory as a skill toolbox that contains a recipe that, if followed, will produce a successful bike ride, baseball swing, or the donning of an SCSR. Interestingly, these recipes operate largely out- side of conscious awareness. This makes it hard for a person to articulate procedural memory. If a person does not think about the specific steps of performing a task, reporting these steps to someone else can be difficult. Thus, procedural memory needs to be assessed by demonstration rather than by verbal report. Having adequate procedural memory for example on how to don an SCSR helps ensure that miners can put these devices on flawlessly even when their working memory is impaired. Another way to characterize the different types of thinking that occur at various skill levels is the “skill, rule, knowledge” approach (Rasmussen, 1983; see also Reason, 1990). The phrases skill, rule, and knowledge broadly characterize the degree of conscious control a person has over what he is doing. For instance, knowledge refers to an activity where a high degree of conscious attention needs to be used to make decisions or perform an activ- ity. This might be the case when a new miner initially learns to don an SCSR. With practice, however, this activity should ideally progress to a rule and then a skill where it can be completed largely outside of conscious control. This classification can be useful to help diagnosis errors. For instance,

OCR for page 63
74 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES an error in donning an SCSR that occurs because the miner automatically skips a step is quite different from not knowing the steps in the first place. By understanding different forms of errors, training practices can be de- veloped to target specific mistakes (see Reason, 1990). This classification model can also be used to help determine when externalized information about completing particular skills—such as checklists or acronyms (see Gawande, 2009)—may be most effective (e.g., if miners repeatedly skip steps, an acronym that includes all the steps could be useful). In summary, the goal is to train miners to a level where they can use necessary breathing apparatus automatically, even though they hardly ever use it. In Chapter 6, we talk in detail about what decision science research says about how to train procedural memory. One theme is the idea that repetition in itself does not ensure adequate proficiency. Rather, mastery, or a demonstration of that proficiency is needed. Miners also have to know when to use breathing apparatus. This needs to be trained to automaticity so that a miner, in the moment, does not have to make decisions when working memory might be compromised (either by stress, CO, or both). For habitual reaction, the miner has to have thor- ough expectations for what donning and using an SCSR is like, and the miner has to trust in the equipment. The miner also has to have experience problem solving on the fly so that dealing with unexpected events also be- comes second nature. Research demonstrates that when tasks have become proceduralized, people have thinking and reasoning resources left to devote to other issues (Beilock et al., 2002). Proceduralizing the components of donning breathing apparatus, and when to don it, leaves valuable cognitive resources necessary to solve unexpected problems in the moment. Miners also have to have adequate knowledge about how other safety technology in the mines work. This includes gas-monitoring devices, com- munication systems, lifelines, and refuge chambers. One way to acquire this type of knowledge is through emergency drills and protocols that spell out, in a step-by-step way, all the information about the mine that might aid in self-escape. Finally, it is important to note that miners also need to be trained in terms of the limitations of the technology they use. They need to know what signs to look for if their equipment is not working or if it needs to be replaced. A thorough understanding of the limitations of the technology and equipment will help prepare the miners to make optimal decisions in an emergency. Knowledge of the Mine Well-practiced primary and secondary escape routes are important for successful self-escape in the event of a mine emergency. Ideally, miners

OCR for page 63
DECISION MAKING 75 should have memorized how to get to an escape route such that they can walk out of a mine in situations where there is limited visibility or in situa- tions where stressful conditions make reasoning or navigating difficult. This knowledge is especially important in situations where the escapeway map on the section is not visible. In addition to rote knowledge of escapeways, it is also important for miners to have detailed knowledge of the spatial layout of the mine as a whole—otherwise known as a cognitive map of the mine. A cognitive map (or mental map) is an internal memory representation of the layout of the mine (Tolman, 1932). Cognitive maps allow a person to visualize the layout of a particular place in one’s “mind’s eye.” Importantly, a cognitive map preserves spatial relations and distances from one landmark to another (Kosslyn, 1994) and thus can play an essential role in helping miners use a landmark to determine the best route for exiting a mine in the event of an emergency. Cognitive maps should not be limited to primary escapeways but should also include basic knowledge of the ventilation system (and how it could change during a mining emergency), caches for breathing appara- tus, lifelines, communication systems, and refuge chambers. Miners should also know how to use the environment to find information needed to self- escape (e.g., use lifelines to determine the location of cache). Cognitive neuroscience research has determined that cognitive maps are derived using visual imagery and many of the same visual processing areas in the brain involved in actually perceiving information in the world are used when people invoke visual images (Kosslyn, 1994). Knowing that visual images share many features with perception lends insight into how mental maps of the mine can be committed to memory. Specifically, miners should not just be given verbal or written information about the layout of the mine but should actually use visual information (maps) to memorize im- portant information. Research also shows that movement through an envi- ronment can help people understand distances and spatial layouts (Burgess, 2006) and that active exploring and having to make decisions about which direction to go in a training situation (Bjork, in press), as opposed to just following someone else out of the mine during training, can also be benefi- cial for learning spatial layouts. Requiring individual and groups of miners to walk escape routes and make decisions about possible paths to safety in training exercises will likely be beneficial for miners developing a thorough understanding of the mine layout. Locations of caches, refuge chambers, and other key places can serve as important landmarks, providing miners with information about where they are in the mine. Route knowledge is thought to develop by registering one’s actions with a set of landmarks in the environment (Siegel and White, 1975). Explicitly teaching miners to think about mine landmarks with re- spect to particular ways out of the mine could prove beneficial for learning

OCR for page 63
76 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES the layout of the mine and for coming up with novel escape paths in the event that self-escape in a mine emergency necessitates changes from prac- ticed escape routes. Landmarks are a way to “off-load” navigation onto the environment (Waller and Lippa, 2007)—as long as these landmarks do not move. For instance, if a miner is aware that a belt line (a landmark) leads out of the mine, the miner can follow the belt line in the event of an emergency with limited visibility. This form of cognitive off-loading may be especially important in stressful situations where effective decision making, planning, or navigation abilities are stunted. Following a belt line or some other stable external landmark limits the need to navigate and reason on one’s own. The lifeline represents one form of cognitive off-loading already in place. Miners can use the lifelines to determine the locations of the near- est escapeway and SCSR cache. Knowledge of What to Do to Self-Escape Successful self-escape in an underground coal mine emergency involves (a) detection, (b) assessment, and (c) escape phases (see Figure 1-1). Detection involves developing conceptual knowledge of common prob- lems indicators. This is akin to how experts build up a rich semantic knowledge base in their domain of expertise (Chi et al., 1981; Lesgold et al., 1988). This knowledge allows miners to classify the problem appro- priately (Chase and Simon, 1973; Ericsson and Polson, 1988) and then to assess the problem, which includes identifying possible solutions. Finally, the escape phase involves the development of “if, then” rules. If a particular scenario occurs, then take a particular course of action. These “if, then” rules can also be considered as procedural knowledge that is enacted fairly automatically once the problem has been identified (Anderson, 1993). As an example, in the physics work mentioned above (Chi et al, 1981), it was found that expert physicists actually spent more time than novices analyz- ing a problem in order to decide what kind of problem it was but less time actually solving the problem. Once experts had categorized the problem, they automatically activated the procedural knowledge needed to solve it and solved it very quickly. These findings are consistent with research on expertise showing that the first option experts generate is usually the best one (Klein et al., 1995; Gigerenzer and Goldstein, 1996; Johnson and Raab, 2003). Finally, these findings are consistent with the idea that sometimes the best action is to pause before a decision is made (Kowalski-Trakofler et al., 2010). Specifi- cally, in an escape situation, it is important for miners to make sure they are aware of all the available information before they act. Making sure that all the available information is collected and used in the decision process is especially important for group leaders in emergency situations. Research

OCR for page 63
DECISION MAKING 77 has shown that pausing to assess a situation and gather new information can allow individuals to come up with the most appropriate response or see a situation in a new way (Wiley, 1998). A basic premise of most human-systems integration approaches is that changes in one part of the system can have an impact on another part of the system. For instance, organizational demands regarding productivity can impact a miner’s bias for determining whether some environmental indica- tor is a sign of a real emergency that requires self-escape is a false alarm. Together, the different parts of a system can often serve to prevent weak- nesses in one part, but sometimes these weaknesses align and adverse events occur. As talked about more in Chapter 5, this aligning of weaknesses is often referred to as the “Swiss cheese model.” The idea is that when holes in different parts of a system line up, unanticipated adverse events can occur (see Reason et al., 2001). Anticipatory Thinking and Heuristics Successful self-escape also involves flexible or anticipatory thinking. Anticipatory thinking is the process of imagining unexpected events and how they may affect plans and practices (Klein et al., 2010). It is a hall- mark of expertise. For instance, expert chess masters are able to plan out several moves ahead in a game situation, and down several possible move trees, to determine whether a particular move will be successful (Chase and Simon, 1973). Importantly, anticipatory thinking is not mere prediction but involves actively interpreting the environment for information that might change a potential course of action. For instance, it has been shown that expert drivers constantly scan the environment for possible hazards in a way that novices do not (Pradhan et al., 2005). Anticipatory thinking allows miners to adapt to changing emergency situations by understanding the consequences of potential decisions and how they need to be altered in the event of changing factors in the environ- ment. It also allows them to adapt to a situation in which several factors come together to lead to unpredictable consequences. Anticipatory think- ing, fueled by expertise in self-escape, may also help miners avoid common mistakes that tend to happen in stressful situations when working memory is compromised. Thought patterns known as heuristics are short-cut strategies for solv- ing problems, which can be useful when decision time is short and reason- ing compromised (Sternberg, 2003). Such short-cuts can be especially useful in situations where the decision maker is dealing with large amounts of in- formation. For example, if a miner has a heuristic for donning an SCSR that smoke = donning, then there does not need to be time taken or cognitive resources spent on considering the pros and cons of donning—especially

OCR for page 63
78 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES when such time could be used to gather important information about the emergency situation at hand. However, some heuristics can lead to biases and ultimately poor decisions. These potential mistakes include, but are not limited to, (a) sunk costs, (b) backup avoidance, and (c) confirmation bias. Sunk costs occur when a person follows through on a decision initially made even if there are signs that this decision should be reevaluated or changed (Thaler, 2000). Making a commitment pushes people to resist revaluation. Backup avoidance is when people do not want to consider an option that will take them further away from their goal at first—even though it may be the best option (Anderson, 2005). This avoidance may have oc- curred during the Aracoma Alma mine fire where miners went forward rather than backward in an attempt to escape (U.S. Department of Labor, 2007b). Perhaps considering going backwards to avoid smoke would have been beneficial, similar to how airline attendants routinely urge passengers to recognize that, in the event of an emergency, the nearest exit may be behind them. Confirmation bias involves looking for information that confirms the story a person has built instead of looking for information that might dis- confirm it (Galotti, 2008). People have a tendency to want to search out meaningful patterns and make sense of experiences and thus they look for information to confirm their initial predictions and tend to ignore factors that could disconfirm it. This means that people may be less likely to pay attention to the environment cues that do not confirm initial assumptions and, as a result, less likely to update erroneous assumptions. And, in many domains, novices tend to do this more than experts (Lesgold et al., 1988). A related idea is illusory correlations, where two events occurring together in time are seen as causally connected even though they are independent (Chapman and Chapman, 1967). Building on these sorts of mistakes in training and educating miners that they may occur can be a powerful way to create the knowledge they need to effectively self-escape in a mining emergency. COMMUNICATION Communication is at the heart of behavioral elements that are funda- mental to self-escape, such as organizing, gathering information, decision making, creating group cohesion, providing guidance, maintaining mo- tivation, and informing and directing effort. This section discusses com- munication between escaping miners, and communication between miners underground and key support personnel on the surface.

OCR for page 63
DECISION MAKING 79 Between Miners As noted in Chapter 3, most escapes occur in groups with miners col- lecting together to move to a place of safety. Sometimes the group repre- sents an intact work team or section crew, but in other situations an escape group is formed by individual miners with varying roles who happen to be nearby at the time of the emergency. Within any group of miners there could be a wide range of experience, expertise, knowledge, and ideas. These are resources, held by members of the escaping group, which should be mobilized to solve the escape problem. In situations where SCSRs are worn or verbal communications is oth- erwise prevented, communication between miners is reduced to rudimen- tary, nonverbal signals and/or writing notes. The mining community has developed a series of hand and headlamp signals (Kosmoski et al., 2012). This approach seems adequate for issuing commands, such as evacuate, go this way, slow down, yes, no, etc.; however, it cannot support questions, detailed statements of information, explanations, or any notion of a con- versation. Although this may be marginally useful for a designated leader, it leaves a poor set of options for followers. Miners are limited in their ability to ask questions, relay a particular piece of important information, or report any physical failings. If a miner believes that the group is going in the wrong direction, the choices are (a) keep silent and moving in what one believes to be the wrong direction, or (b) remove an SCSR to talk and risk toxic inhalation, (c) leave group and turn in another direction alone without explanation, or (d) try to communicate with gestures, taps, and mumbles. The miner might be able to communicate on the tagline1 by shak- ing it or by grunting out loud, but this type of communication is lacking in terms of the richness of information it can convey. A nonverbal scenario forces an escape group to rely almost exclusively upon the knowledge and decisions of only one person. Recent research has demonstrated that groups of individuals working together make the best de- cisions when there is a high level of collective intelligence (known as the “c factor” [Woolley et al., 2010]). Collective intelligence is not strongly related with the maximum intelligence or knowledge of individual group members (i.e., with what one person knows) but rather with the ability of a group to communicate (e.g., take turn in conversations, exchange information). Thus, the ability to communicate within a group of miners in an emergency situation seems highly beneficial for successful self-escape—especially when there are changing circumstances that require the reevaluation of initially chosen options. 1  tagline is a long heavy-duty rope with tethers spaced at even intervals, designed to link A members of a mine crew together in the event of an emergency, particularly in dense smoke and little or no lighting.

OCR for page 63
80 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES When verbal communication and ongoing exchange of information is possible between miners, the members of the escape group are able to participate in all of the fundamental behavioral elements of self-escape discussed earlier. Miners will differ in experience and knowledge, but the resources that exist within the group now can be mobilized. Verbal com- munication enables miners to contribute information about a fire or other hazards and the locations and status of personnel and to suggest courses of action, weigh options, ask questions, give opinions, sound objections and explain them, and so forth. With verbal ability, what would otherwise be a collection of individuals now has the potential of becoming an effective escape team or group. Interventions designed to improve team effectiveness, such as leadership and followership training, can now be useful. Improving team communication is critical given that people have an egocentric bias to think others understand what they have communicated, even when others do not (Chang et al., 2010). Training that allows miners to develop accurate communication strategies—verbal and nonverbal—would be an essential component of successful self-escape. Between Miners and Surface Personnel The surface communication center, the responsible person, and the im- mediate support team can be significant resources to miners, particularly when verbal communication is possible.2 As with the case of communica- tion between miners underground, any constraints on communications with the surface limit important coordination and information exchange between the personnel on the surface and underground escape groups. The surface communication center can play an integral part in the exchange of critical information from first alert to any time that the escape groups are able make contact. Surface personnel obtain information from the miners and provide other information back to them. The information obtained by the responsible person is routed to other surface personnel for decision making with respect to locations, firefighting, aiding escape, and rescue. It will include such information as locations of the fire and the miners, health status of the miners, availability and usability of breathing apparatus, presence and density of smoke, and miners’ intentions. Some of the information obtained from one escape group is also subsequently routed back down to other groups when they contact the surface for information. The information obtained from the surface by the miners is used to make 2  exting T possibilities exist in the absence of verbal communication that may permit limited text communication between surface communication centers and miners underground. How- ever, the number of mines with this capability is relatively few, and the speed and depth of communication is limited.

OCR for page 63
DECISION MAKING 81 decisions, such as route choice, wayfinding, the status of other miners, and entering or not entering a refuge. Given that the responsible person and team may need to simultane- ously obtain and provide valid information to disparate groups of miners underground and to other personnel on the surface under stress and time pressure, it raises issues related to communication, such as: How should the responsible person support team and its task be structured? How many people are necessary to do the job? How should they divide their roles? What training should be provided to them? As these issues are addressed, modifications to current arrangements should be directed at clarifying roles and simplifying the significant communication responsibilities that the re- sponsible person carries. IDENTIFYING SELF-ESCAPE COMPETENCIES Decision science research has provided insight on the types of infor- mation miners might miss and the common mistakes that may be made in emergency situations (see Box 4-1). This research has helped identify cognitive competencies necessary for the self-escape task and as such can inform the development of emergency protocols and training procedures. This section briefly discusses five critical competencies—detecting hazards, using equipment and technology, wayfinding, understanding stress, and team functioning. We note, however, that this list of competencies is drawn from the discussion in this chapter but is notably incomplete. As discussed in Chapter 3, a full critical incidents analysis and task analysis would be necessary before a complete list of competencies could be identified. Detecting Hazards: Miners have to constantly draw distinctions between routine hazards and those that require self-escape. To do this, miners must have knowledge of environmental conditions that require self- escape and/or use of personal protection equipment. Miners also have to understand how biases (e.g., an organizational culture that implicitly discourages false alarms or their own lack of trust in safety equipment) might impact their decision to label something as an emergency. Min- ers need to develop a rich knowledge base that allows them to auto- matically know which environmental cues mean they should don their breathing apparatus and what self-escape procedures should be enacted. Using Equipment and Technology: Miners need to be able to automati- cally (without thinking in detail) don breathing apparatus and switch from one to another. They also have to have very clear expectations for what donning is like. They have to be able to fluently use other

OCR for page 63
82 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES BOX 4-1 Examples of Psychological Factors That Can Affect Effective Self-Escape Optimism bias—The human bias to initially assume that nothing is wrong. In the context of a mining emergency, this may involve ignoring initial signs of fire or roof falls. Cultural or organizational bias not to false alarm—Tacit pressure from an organi- zation not to behave as if self-escape is required (e.g., not donning self-contained self-rescuer [SCSR] when environmental cues such as smoke indicate otherwise because a miner is hesitant to use an expensive piece of equipment). Compromised thinking and reasoning under stress—The decrease in a person’s cognitive capacity, the ability to think and reason systematically, which is often compromised under stress. Recognizing this can aid decision making. Emotion-driven decision processes—A person’s tendency to allow emotions to dictate decisions. May result in putting additional people in danger. Backup avoidance—The tendency to not want to go away from one’s goal. This tendency may result in miners not considering escape routes that initially take them farther from a place of safety but are ultimately the best choice. Confirmation bias—The tendency to only look for information that confirms what one believes (e.g., about the cause of an emergency situation) and thus not up- date one’s notion of what has happened and what needs to be done to effectively self-escape. Egocentric communication bias—One’s bias to assume that others understand what one has said, even when they have not, which may increase in times of stress. Sunk costs bias—The tendency to continue following through on a decision initially made even if there are signs that the decision should be reevaluated or changed. technology relevant for self-escape (these include, but are not limited to, communication devices and gas monitors). Wayfinding: Miners need to have adequate awareness of their environ- ment. This knowledge includes mental maps of how to get to escape- ways, how landmarks can help them determine where they are in the

OCR for page 63
DECISION MAKING 83 mine, and how they should travel out in addition to utilizing current lifeline symbols. Understanding Stress: Miners need to have awareness of how stress impacts decision making (e.g., how the brain and body changes in stressful situations) and the types of decision-making mistakes and potential pitfalls that are likely to occur in mine emergency situations. Team Functioning: The ability to function as an effective member within a team is also a fundamental competency that emerges in self- escape. Although the ability to self-escape alone has to be supported, most escapes occur in work groups. To escape, these work groups must transform themselves into teams in which members have roles and responsibilities, share the common goal of escape, share a common mental model or understanding of how an escape team should function, and work to enable the team to be successful. Toward this end, a team member must understand the various ways in which one can contribute (e.g., providing information to leader or group), when to communicate and when to listen or encourage others to speak (e.g., detecting situ- ational cues), and when to be a leader and when to be a follower (e.g., delegating and accepting tasks). With respect to working with surface personnel, communication skills are obviously important (e.g., passing along facts and flagging opinions as such). RECOMMENDATION The findings from research in the field of decision science, which can broadly be defined as the investigation of decision processes and commu- nication strategies within and across people, is increasingly recognized as important for understanding human behavior across a variety of fields. To effectively self-escape in the event of a mine emergency, miners need to have more than knowledge of their equipment and surroundings; they must also have the psychological tools to make effective decisions and communicate successfully. Decision science research helps identify common thinking and reasoning pitfalls that can occur in stressful situations (see Box 4-1) and also informs the training that miners take part in as a means to ensure suc- cessful self-escape. RECOMMENDATION 5: The National Institute for Occupational Safety and Health should use current decision science research to in- form development of self-escape training, protocols, and materials for training for effective decision making during a mine emergency. Miners

OCR for page 63
84 IMPROVING SELF-ESCAPE FROM UNDERGROUND COAL MINES and mine operators should be knowledgeable of typical warning sig- nals and able to determine if a true emergency exists and decide how to respond appropriately. All miners should be trained using standard protocols developed for predictable components of self-escape. This will allow miners to devote adequate attention to unexpected events and enhance situational awareness.