هوشیاری ناممکن: سعی و کوشش، حواس پرتی و خیال بافی منجر به شکست در یک کار نظارت عملی

Abstract In laboratory studies of vigilance, participants watch for unusual events in a “sit and stare” fashion as their performance typically declines over time. But watch keepers in practical settings seldom approach monitoring in such simplistic ways and controlled environments. We observed airline pilots performing routine monitoring duties in the cockpit. Unlike laboratory studies, pilots’ monitoring did not deteriorate amidst prolonged vigils. Monitoring was frequently interrupted by other pop-up tasks and misses followed. However, when free from these distractions, pilots reported copious mind wandering. Pilots often confined their mind wandering to times in which their monitoring performance would not conspicuously suffer. But when no convenient times were available, pilots mind wandered anyway and misses ensued. Real-world monitors may be caught between a continuous vigilance approach that is doomed to fail, a dynamic environment that cannot be fully controlled, and what may be an irresistible urge to let one’s thoughts drift.

1. Introduction In 1972, the crew of Eastern Air Lines Flight 401 became distracted when an indicator light designed to verify that the airplane’s landing gear was down and locked failed to illuminate. As the crew worked together to resolve the problem, which in cruel irony turned out to be only a burned-out light bulb, the airplane continued a slow and unnoticed descent into the Florida Everglades (NTSB, 1973). In response to this landmark accident, airlines redefined the roles of the two pilots in the cockpit. From then on, one pilot (called the pilot flying) would assume primary responsibility for operating the controls of the airplane, while the other pilot (the pilot monitoring) would serve as the “second set of eyes” in the cockpit, occasionally assisting the pilot flying in his or her duties but primarily performing the job of keeping watch for anything amiss. Moreover, airline companies devised explicit procedures to be used by monitoring pilots. Monitoring pilots were formally tasked with making verbal callouts as airplanes climbed or descended to or away from assigned altitudes in order to raise awareness among the crew about what the airplane was doing. But despite these remedial steps, accidents related to lapses in monitoring continued to occur (NTSB, 1994). As recently as 2013, yet another crash occurred in San Francisco after the flight crew failed to notice that the airplane had slowed to an unsafe speed during approach (NTSB, 2014). An unsettling detail about the San Francisco accident was that, unlike most flights in which two pilots sit in the cockpit, this flight had four. None of the four pilots noticed the deteriorating airspeed that was shown on a simplistic cockpit instrument and positioned in plain view directly in front of them. The persistence of airline crashes related to monitoring failures, despite the attention given to most every conceivable aspect of the problem, lead us to wonder if there is something fundamentally wrong with the very idea of placing a human in a supervisory role, however well-trained and experienced they may be. Do our troubles begin when we so much as attempt to watch over the shoulder of someone or something else while it works? 1.1. Why vigilance sometimes fails Upon first glance, the problem of keeping watch for things gone amiss in an environment such as an airline cockpit seems simple enough. Unfortunately, countless studies have shown us that the process of paying attention while watching others work can go wrong in several ways. 1.1.1. Depletion Thousands of laboratory experiments have demonstrated what happens when participants are asked to monitor a stimulus or process in a prolonged “sit and stare” fashion. What is invariably found is that a decline in performance eventually sets in (Mackworth, 1948 and Teichner, 1972). Many argue that the cause of this vigilance decrement is the depletion of cognitive resources needed to perform what is an arduous, stressful, high-workload task (Dillard et al., 2015, Helton and Warm, 2008 and Warm et al., 2008). As participants work to maintain their focus on whatever they have been assigned to monitor, their stores of cognitive energy are depleted faster than they can be replenished. 1.1.2. External distractions Early researchers such as Mackie (1987) were quick to point out that, in real-life monitoring tasks, outside the guarded door of the research laboratory, people seldom have the opportunity to sit and stare for extended periods of time. For example, monitoring pilots are periodically asked to answer radio calls from air traffic control or to read checklists. In separate studies of lifeguards and prison guards, Harrell (1999) and Tickner, Poulton, Copeman, and Simmons (1972) found that guards’ scanning was also frequently interrupted by other duties such as responding to rule violations that they are tasked with monitoring. The problem with these momentary distractions from the monitoring task is that they too can lead to monitoring misses. Studies of the deleterious effects of having to divide our attention between more than one task go beyond demonstrating that we are likely to miss something that lies to the left when our attention is momentarily diverted to the right. It has been demonstrated that we often “look but do not see” when cognitive resources needed to meaningfully process what we are monitoring are tied up by another task such as talking or listening (Strayer & Johnston, 2001). Einstein, McDaniel, Pagan, and Dismukes (2003) demonstrated that even momentary attentional diversions often lead to forgotten intentions associated with a monitoring task. 1.1.3. Internal distractions Other researchers have shown us that, while people work, they sometimes turn their thoughts to matters that are unrelated to the task at hand, and that during these mental excursions, things can slip by unnoticed (Smallwood & Schooler, 2006). Some researchers argue that mind wandering is the greater threat during a mundane vigilance task (Manly, Robertson, Galloway, & Hawkins, 1999). They argue that mind wandering gradually encroaches upon us, often without our awareness, when we perform monotonous or boring tasks over prolonged periods of time. Ariga and Lleras (2011) suggest that we may knowingly engage in mind wandering when we wish to take mental breaks from a tiring monitoring task. But Kurzban, Duckworth, Kable, and Myers (2013), and Thomson, Besner, and Smilek (2015) have yet a different take on mind wandering. These researchers suggest that we turn to mind wandering when we feel that our attentional resources might be more profitably directed elsewhere. After assessing the frequency at which events of interest might occur during a monitoring task, people may reallocate their limited attentional resources based on the likely reward for monitoring and the opportunity costs associated with not doing or thinking about something else. Casner and Schooler (2014) have already demonstrated that the successful use of highly reliable cockpit automation is associated with frequent reports of task-unrelated thoughts among airline pilots. 1.2. Is vigilance impossible? These three reasons for which monitoring is suspected to fail, that have been so often studied independently in laboratory settings, combine to offer a rather bleak outlook for those who monitor in real-world settings. It seems that if we set out to actively monitor for long stretches of time, our performance predictably slumps. If we choose to (or are required to) perform other tasks that may arise, we get distracted and miss things for another reason. And if we allow our thoughts to drift onto other matters to take breaks, because we think our attention is not needed, because we find mind wandering irresistible, or simply because we think we can get away with it, we miss things for yet another reason. Whether we are diligent, distracted, or daydreaming, is monitoring doomed to fail? 1.3. An observational study of monitoring in the airline cockpit We designed an observational study of airline pilots to see if this seemingly cheerless combination of psychological factors could help explain the persistent monitoring problems we see in the airline cockpit. If our theory about the inevitability of monitoring failures is accurate, we should expect monitoring failures to copiously occur, even when highly-experienced airline pilots perform under the watchful eyes of experimenters. And if we do witness a sample of cockpit monitoring lapses, the three reasons for monitoring failures should be able to account for at least a substantial proportion of them. We asked sixteen airline pilots to act as pilot monitoring during routine flight operations in a full-motion flight simulator certified for airline training and testing. As our sole dependent measure, we noted whether or not pilots were successful in making the required altitude callouts that are central to the pilot monitoring job. We chose the altitude callout task for three reasons: (1) because it is one that is performed routinely during every flight; (2) because it is considered to be an indicator of overall monitoring performance by airline training departments (Sumwalt, 2003); and (3) because lapses in altitude awareness have been associated with numerous incidents and accidents (NTSB, 1994 and Palmer et al., 1991). We asked pilots to monitor and make altitude callouts during two different phases of flight that differed in important ways. During the arrival phase, airplanes follow a published procedure that spells out in advance the series of altitudes that the airplane must cross. For the monitoring pilot, as soon as they monitor and call out one altitude, responsibility to monitor and call out the next one is immediately upon them. During an arrival, altitude monitoring is a continuous process that offers the monitoring pilot no breaks from the altitude callout task. During the approach phase, the environment is much less predictable. During the approach phase, the airplanes have arrived to the busy airspace that surrounds the airport, where air traffic control must concoct off-the-cuff plans for keeping airplanes separated. Airplanes will field directives from air traffic control who ask them to climb or descend. Once a climb or descent has been accomplished, the flight crew must wait for the air traffic controller to issue their next assignment. Thus, the approach phase offers the monitoring pilot breaks from the altitude callout task. To assess the effects of internal distractions on success at the callout task, as pilots worked, we verbally probed them about the current focus of their thoughts. Pilots were asked to report whether they were thinking about something related to or unrelated to the flight. The verbal thought probing technique is an adaptation of a technique introduced by Singer (1966) and has since been shown to be minimally intrusive (Ericsson and Simon, 1980) and correlate well with physiological measures of task-unrelated thought (Schooler et al., 2011, Smallwood et al., 2011 and Smilek et al., 2010). Most importantly, the thought probe technique been successfully used in several recent studies (Casner et al., 2014 and Casner and Schooler, 2014) that examined mind wandering among airline pilots. These studies have demonstrated pilots’ routine willingness to acknowledge off-task thoughts when queried, and a relationship (at least in some situations) between such thoughts and performance detriments. To assess the effects of external distractions on success at the callout task, at the time that each thought probe was given and again when each altitude callout was due, we noted whether or not the monitoring pilot was engaged in any non-monitoring activity such as communicating with air traffic control or the pilot flying, or whether they were interacting with any of the aircraft systems. To explore the association between mind wandering and the monitoring role, we asked pilots to additionally act as the pilot flying during a portion of the study while we continued to probe them for their thoughts.

. Conclusion We studied monitoring performance in an important vigilance situation: airline pilots serving in a monitoring role. A first striking result was the number of monitoring lapses we observed while pilots performed a routine cockpit monitoring task. Pilots missed 25% of the routine altitude callouts that they were tasked with making. Unlike laboratory experiments in which participants’ performance gradually deteriorates as they endure long periods of watch, our pilots avoided long vigils and escaped their deleterious effects. One way in which our pilots avoided long periods of watch was to engage in other cockpit activities that are designed into the job of pilot monitoring. Unfortunately, while these breaks from the monitoring task may have helped pilots avoid one sort of failure (i.e., declines in vigilance over time) they seemed to have led them directly to another: the distractions introduced by these non-monitoring tasks were linked to their own monitoring misses. Another way pilots avoided long vigils was to engage in mind wandering: an activity over which pilots may have exercised notable conscious control. But when pilots could not find convenient times to mind wander, mind wandering sometimes led pilots to yet another source of monitoring failure. These results support our theory that whether pilots remain attentive or not, the task of monitoring may offer few paths to success. Engaging in diligent monitoring, mixing up the monitoring task with other sorts of activities, or taking mental breaks all seem to lead to monitoring lapses. The implication that monitoring lapses are inevitable is hard to accept given that we rely on human monitors to keep us safe in a great many situations. While it is unlikely that we will eliminate the problems associated with long uninterrupted watches, our findings suggest that the two ways of avoiding them that we have studied here (internal and external distractions) might be better managed. Pilots made no apparent attempt to avoid the deleterious effects of engaging in non-monitoring tasks when an altitude callout was soon due. It may be that pilots were often not free to postpone activities performed at the request of others, or that their confidence in their ability to multitask far outstrips their real ability to do so (Finley, Benjamin, & McCarley, 2014). We might try educating pilots about the “multitasking myth” (Loukopoulos, Dismukes, & Barshi, 2009) or designing a reminder system to help pilots deal with interrupted tasks (Einstein et al., 2003). Palmer and Degani (1991) have explored the idea of using electronic checklists to improve monitoring performance but discovered that adding more automation to the monitoring problem might further lessen pilots’ engagement in the task. Pilots devoted 43% of their available monitoring time to task-unrelated thought, and we observed many instances of mind wandering in the vicinity of monitoring misses. We might consider ways of reducing the overall amount of mind wandering. Less practically, MacLean et al. (2010) have demonstrated the benefits of meditation and mindfulness training on subjects’ ability to perform a sustained attention task. Founded on the notion that a busy mind is an engaged mind, deBettencourt, Cohen, Lee, Norman, and Turk-Browne (2015) have demonstrated improved attention through modulating task difficulty when attentional lapses were detected using whole-brain neuroimaging. More practically, Sumwalt (2003) has suggested that monitoring performance might be improved by tasking pilots with actively monitoring a more comprehensive set of flight parameters. However, despite the existing formal requirement to monitor altitude, our pilots still mind wandered to great extent and sometimes missed callouts. That our pilots may have demonstrated more control over their mind wandering than participants in other studies (Smallwood & Schooler, 2006) raises the question of whether or not the ability to minimize mind wandering or confine it to more appropriate times is a skill that can be selected, acquired, or even taught. By way of selection, Kane et al. (2007) found that subjects with higher working memory capacity were more successful in maintaining on-task thoughts when concentration and effort was needed. By way of training, we might consider using techniques to manage mind wandering similar to those used to manage fatigue (Rosekind et al., 1995). In closing, we revisit the idea that monitoring is a task that requires the same tight-looped activity to be performed continuously, and where penalties for lapses can be severe. To accomplish this task, we often use a single human being who has limited attentional resources, who is easily interrupted by events both internal and external, and yet whose overall success may depend on all of that. In critical situations where lives are at stake, perhaps we do not have a good match here. But for as long as we continue to place humans in the monitoring role: in airline cockpits, beside swimming pools and baggage scanners, atop prison walls, and inside operating rooms, we should be grateful for the safety record we have today.