Control in Work Groups – Monitoring as Collaborative Task Management
Collaboration on the flight deck requires individuals to reliably fulfil their part of the joint task. The basic framework of activity is provided by procedures but collaboration also involves activity at a metacognitive level in that individuals’ need to independently exercise oversight. Crew members coordinate action with others but also validate progress in order to decide whether to intervene or modify activity as required. We call this process ‘monitoring’. Failure to ‘monitor’ is now well- established in the lexicon of accident report writing. The UK Civil Aviation Authority (2013), the Flight Safety Foundation (2014) and IATA (2016) have all produced guidance material on improving PM skills and it has even been suggested that ‘training in monitoring’ be included in the revised European airline recurrent training guidance (EASA, 2018). Dismukes and Berman (2010) offer a cautionary note: whatever you do to improve monitoring, the process will fail. Monitoring, or routine scanning, was shown to be the least effective method of error detection in Chapter 6. This section positions ‘monitoring’ as a social process at the heart of collaboration.
Historically, the operation of an aircraft was considered to be the job of the aircraft commander who sat in the left-hand seat. With the advent of two-pilot cockpits on commercial passenger aircraft, the right seat was occupied by an assisting, albeit qualified, pilot. This division of responsibilities is manifested in the layout of early generation cockpits where the left seat had a full array of instrumentation while the right seat pilot typically had a sub-set of displays but also had many of the controls for aircraft systems. The body in the right seat was clearly intended to be in the supporting role.
The change in operating philosophy, such that both pilots took turns to operate the controls, required a change in nomenclature. Various terms were used to differentiate the pilot actively in control of the aircraft from the pilot acting in support. Federal Aviation Administration (USA) (FAA) Advisory Circular 120-71A (FAA, 2003) formally proposed the adoption of PF and PM to distinguish between the two roles. The Circular also introduced an appendix on crew monitoring and cross -checking but addressed the issue through the development of SOPs rather than explaining what monitoring is and how it should be undertaken. A common understanding of the PM (or ‘pilot non-flying’ (PNF)) role is to ‘monitor the flight management and aircraft control actions of the PF and carry out support duties such as communications and checklist reading. The Operations Manual will specify fully the roles for the PF and PM/PNF, but one of the most important aspects of the duties of any PM/PNF is the cross-check of the actions of PF (Skybrary, 2018)’. In order to clarify how pilots view the concept of monitoring, pilots attending an annual crew resource management refresher course were asked to discuss ‘monitoring’. Nearly 2000 pilots participated over a 12-month period. Pilots were divided into two groups, captains and first officer/second officers (FO/SOs), and set a number of tasks designed to clarify and elaborate on the concept.
One interesting insight was that ‘one size does not fit all’, the implication being that monitoring, like flying, is an adaptive skill and needs to be modified to suit the monitoring load represented by the nature of the operation and the skill set of the PF. This is further supported by a cluster of comments relating to differences in experience levels within and between the two groups. For example, it was observed that training captains often lacked the routine familiarity with line operations and some domiciled (based) pilots were unfamiliar with clusters of destinations they rarely operate to. Interestingly, trainers themselves freely admitted that they were aware of a lack of ‘sharpness’ in their skills and many suggested that they needed breaks from the training programme just to keep their line currency up to standard. This increased the monitoring load on the FO. These observations reinforce other comments about differences in competence and experience as being limiting factors in collaboration.
It was clear from the outset that the groups did not see the PF/PM roles as being mutually exclusive. As PF you are also monitoring the activity level of the PM to ensure that you get the support you need. As PM you are actively involved in the flying task simply because of procedure design and task allocation. This can interfere with monitoring the bigger picture because it is a cause of distraction. With this proviso, the comments of the groups fall into three main clusters: describing what is monitoring, describing the activity associated w'ith monitoring; elaboration of the role of the monitoring pilot.
Turning, first, to monitoring as a concept, both the captains and the FO/SO groups described monitoring in terms of creating and sharing mental models, keeping the big picture, maintaining awareness and mentally flying the aircraft. So, monitoring is underpinned by accurate mental representations of the intended progress of the aircraft. As such, then, effective monitoring is predicated on sound Level 1 behaviours.
However, clear differences emerged between the two groups that reflect the social dynamic. These differences relate to the way in which the two groups differ in terms of deciding on how progress was to be gauged: in effect, who decides if progress is acceptable. When captains are in the monitoring role, the FO is, of course, the flying pilot. This means that the captain has delegated, to a degree, responsibility for the conduct of flight to the FO. Comments by captains included:
These comments from captains reflect the importance of having a clear plan for the progress of the flight. The final comment also captures the idea of sense-making involving pattern matching between the expected and observed states of the world. The FO/SO group, on the other hand, talked in terms of:
The captain groups’ comments are active, they relate to comparing actual progress against the intended plan. The FO/SO group, on the other hand, used passive terms that reflected a need to verify their own understanding of what they are observing. So, although the group have a common understanding of how monitoring functions (by establishing an accurate overview of the task) they differ in the extent to which they have confidence in the accuracy of their world view. This will degrade an FO/SO’s ability to undertake monitoring. The workshops did not explore differences between FOs in terms of experience but one conclusion that flows from these comments was that competence was based on knowledge structures and, at some point, pilots shifted from an internal view of the work process (‘am I sure I understand what is happening?’) to an external view (‘is what is supposed to happen actually what I’m seeing?’). This first group of comments, then, is an elaboration of the discussion of sense-making in Chapter 5 but in a more applied context. The differences between captain and FO comments reflect how an experience changes the nature of sense-making and immediately points to a potential for breakdown between the two roles.
The second cluster of comments related to the actual task of ‘monitoring’: what is going on when you are monitoring. The views of the captains suggest that monitoring is a set of fixed and variable tasks. The significant fixed tasks were:
In the context of competence, captains stressed the importance of ‘having an idea of what to expect before checking’, which links to the earlier comment about knowledge structures. Importantly, captains also said that the task involved ‘have a comfort zone and know when it has been exceeded’. This comment points to the social dynamic at work. The implication is that impinging on the ‘comfort zone’ will trigger a change in the dynamic: captains will intervene. This is the variable component of monitoring. The FO/SO groups’ view of the task of monitoring offered a much richer description. Initially, they offered a conventional description of the task:
Of interest here is that the comments reflect the subordinate role of ‘assisting’ whereas the captains’ comments were more global in outlook. The next cluster of comments elaborates on the captains’ observation about SOP compliance:
One syndicate referred to monitoring as ‘quality control’ and this function was borne out by several comments that reflect the idea of process control:
Finally, a repeated theme was the need to actively observe the actions of the PF:
Of course, there was no evidence that captains did not do the things on the FO/SO list when they were in the PM role. In all probability, they undertook these actions but did not see them as noteworthy. In fact, one interesting comment was that monitoring ‘happens subconsciously’. This second cluster of comments supports the importance of fundamental procedural knowledge. Collaboration is underpinned by a shared procedural framework which has been sufficiently internalised such that performance is almost automatic. Of course, the challenge is to have mastery of procedures while keeping track of any changes. Maintaining currency is effortful and one reported mechanism for updating procedural knowledge is simply relying on being corrected during recurrent training sessions rather than arriving for training having reviewed the current procedural status. This behaviour suggests that the procedural framework held by some crew is imperfect, drifting towards a generic status until being periodically recalibrated. The gap between the intended framework and that held by individuals might only be small but it is an incremental addition to the broader gap between crew members that reflects differences in competence and experience. I proposed in Chapter 6 that homogenisation was a category of drift at the level of the individual. This discussion of the problem of maintaining procedural currency suggests how homogenisation might start.
These first two categories of comments build on the previous discussions of sense-making and underpinning knowledge. Although both are essential to individual action (Level 1 activity), they create the context for collaboration. The final cluster of comments from the workshops touch on social control of performance. I mentioned earlier that, when the captain is monitoring, the implication is that the task of flying has been delegated to the FO but the social construction of that act of delegation is implicit in the observations by some captains who said that monitoring was:
These comments indicate that ‘delegation’ is an implicitly negotiated transaction. The FO in the PF role has full authority to operate the aircraft, but only up to a point. The view that the FO operates at the discretion of the captain was further revealed by the second cluster of comments:
The tone of the comments is interventionist. Not only does the captain PM exercise oversight but they do so from the perspective of what they consider to be the ‘correct’ method of working. On the other hand, the FO/SO groups were quite blunt in their assessment of the monitoring role:
These views echo the comment of the Learjet FO in our opening case study: ‘my job is to “protect the ego of the captain” and not to speak up unless I was about to die’. These role-related comments flow from the manner in which roles are interpreted, a reflection of role drift, and the way in which power is exercised in the context of the crew. And, of course, they are not unique to aviation. Rather, it is the cost of being human.
All the groups made frequent reference to the need for a ‘shared mental model’. It seems that monitoring is the epitome of collaborative action and to be an effective collaborator requires a clear understanding of:
Monitoring, then, occurs at the intersection of fundamental competence, prescribed work, social dynamics and internal dialogues. Rather than consider ‘monitoring’ to be a metacognitive control process somehow separate from, or additional to, other activity, it seems that it is simply the task of doing work in a team. It suggests that to understanding normal work we need to be cognisant of these multiple dimensions.
The groups were also asked why they felt monitoring fails. There were six key groups of factors that lead to a breakdown. Three were to do with aspects of the task (with some physiological aspects linked to the theme). The other three relate to discretionary behaviour. In the task-related group, familiarity and associated complacency featured strongly in the lists of reasons why monitoring failed and this is borne out by some airlines’ LOSA findings that show that home base is often the location for most errors by the crew. Task overload or under-load, leading to boredom and distraction, was the next most frequently cited causal factor. Stress and fatigue, especially, were also included in this group. Finally, operational issues such as unfamiliar airports, procedural changes, language (strong accents or pace of delivery) and the unexpected (e.g. technical failures, go-arounds) were operational factors that cause problems. This last cluster captures the idea of threats and perturbations.
In terms of discretionary behaviour, the crew dynamic (flight deck gradient, captain’s style), inconsistent application of SOPs and style of communication were the main causes of a monitoring breakdown. In discussions, examples were given of failures to extend the crew concept to augmented crews, that is, when additional crew members were carried. So, for those phases of flight when three or even four pilots were on the flight deck, there was often little or no attempt to formally include the additional crew in the process of flight management. This can also be seen, for example, in smaller aircraft operations. Some types, such as the King Air or early model Cessna Citation jets, are approved for single-pilot operations but the terms of a commercial contract or an insurance requirement might mandate two-crew operations. I once worked with a helicopter operation that used a single-pilot aircraft but required an additional pilot for operations beyond a certain range offshore. No attempt was made to adapt procedures to accommodate the additional pilot. Where there were no clear multi-crew procedures, accidents can, and have, occurred.
I suggest that an emergent property of Level 2 in my hierarchy is efficiency. Effective collaboration will deliver an efficient performance. Sense-making and competence are building blocks of collaboration and will support efficiency but, clearly, the social dynamic is the third pillar.