Desktop version

Home arrow Management

  • Increase font
  • Decrease font


<<   CONTENTS   >>

Decision-Making as Task Management

In the next two sections, I want to look at decision-making in terms of actions associated with the active goal state and then decision-making as adopting a new goal. Having just looked at information processing, the other two components in the cognitive load model, time on task and task switching, relate to the activity associated with the current, active task. Doing work involves individuals engaging with the world and making interventions appropriate to the needs of the task. Because of the inherent variability of normal operations - non-ergodicity - individuals have to adapt their performance within the framework of policies and procedures. Selecting the most appropriate action, therefore, is an act of decision-making. Task management involves making decisions about initiating, sequencing and controlling actions. Task switching is the act of shifting between specific target activities depending upon an assessment of priorities. Crucial in all of this, of course, is the timing of interventions: delay reduces downstream opportunity, which is what we mean by margins being eroded.

Trained and competent individuals have a set of expectations about how a task is to be undertaken based on prior training, the applicable procedures and any relevant experience. As we saw in Chapter 1, these factors are encapsulated in a ‘plan’ or ‘script’ for how the flight will unfold, including the relevant constraint sets that must be satisfied, all stored in LTM. There will be several different scripts for different flight segments and probable common contingencies. Scripts comprise aggregations of schemata relevant to the task. Thus, at a mental level, tasks can be considered as sets of component activities with triggers that initiate an activity and feedback loops that allow for monitoring, all compiled into routines. Tasks will vary in their complexity, the demands they make upon specific skill sets or the competence of individuals, the level of accuracy demanded and the time available for completion. We can identify three criteria for assessing task accomplishment: the quality of task activity, its completeness and its validity (efficacy, or fitness) in relation to goal achievement. Under normal conditions, work involves drawing on the following knowledge structures:

Declarative Knowledge

What to do - what are the steps to the next goal?

Why do it - what is the rationale for this specific course of action?

Process Knowledge

How to do it - what is the procedural activity required to achieve the goal?

Work, of course, is messy and involves enacting bundles of tasks: rarely will a skilled operator be engaged in a single activity. Task elements will be conducted concurrently, in series and interleaved. For example, imagine a light aircraft flying the downwind leg in the traffic pattern. Keeping the aircraft straight and level represents two concurrent, albeit integrated, tasks. Configuring the aircraft for landing (taking flap and lowering the undercarriage) represents tasks conducted in series. Making radio calls is representative of an interleaved activity. This suite of behaviours requires management if the whole performance is to be orchestrated successfully.

Task management activity comprises initiating, monitoring, interrupting, resuming and terminating task elements, all of which are outcomes of decisions made about the current state of progress towards a goal. The scale of task management problems was shown in a study by Chou et al. (1996) who found that 76 out of 324 accident reports (23%) and 231 out of 470 incident reports (49%) contained cockpit task management errors. Task switching is effortful, and most of the problems associated with normal task management stem from poor prioritisation and forgetting. We will look at the error in more detail in the next chapter, but ‘forgetting’ is a specific type of error called a lapse. Initiating new tasks and tracking those tasks that are in play places a burden on cognitive capacity. Deviations from the script, such as changes, loops (repetitions), deferrals or omissions, will increase the cognitive demands, requiring additional attention. That task management is effortful is captured in the concept of workload. Task management is the process of organising w'ork while cognitive load is the controlling of the mental framework and workload is the subjective assessment of the effort involved. The workload is a function of the complete assemblage of activities being undertaken by operators over a unit time and includes both formal activities associated with an immediate task and other informal activities. Social conversation between individuals requires attention and, thus, contributes to workload. The workload can be physical as well as cognitive. One commonly-used tool for assessing workload is the NASA task load index (NASA TLX) (Hart and Staveland, 1988). It comprises six scales that look at different aspects of workload. In an attempt to better understand the cognitive demands of flying departures and approaches, I asked the crew to complete the NASA TLX at the top of climbing and after landing. Table 5.1 shows the mean scores of the cognitive effort for the pilot flying (PF) and the pilot monitoring (PM) The difference between the roles was not statistically significantly, but the difference for the PM between the two stages of the flight was significant. In short, the monitoring role requires more cognitive effort during the arrival phase. We also saw in the last chapter that the threat rate, which

TABLE 5.1 NASA TLX Scores

Top of Climb

After Landing

PF (и = 21)

20.3

22.8

PM (n = 20)

21.5

27.0

is indicative of the rate of perturbations encountered by the crew, adds to workload as measured by the NASA TLX (Aeschbach et al., 2017).

Prioritisation is the process of making decisions about sequencing activity. Colvin (1999) identified three factors associated with prioritisation: status, procedure and value. The status of the task describes the extent to which the operator believes the conduct of the current task is satisfactory. Colvin’s use of ‘status’ is similar to my use in relation to goal states. The factor labelled ‘procedure’ captures the degree to which the operator believes the activity is correct within the overall plan. Value refers to the importance of the task within the current activity and the perceived consequences of not completing the task. Colvin’s ‘value’ maps well onto the concept of efficacy. Prioritisation, of course, may require switching between the concurrent tasks, and there is evidence that the interruption associated with switching impedes task resumption. Furthermore, competing tasks induce forgetting, which is not simply a function of workload. Forgetting is representative of poorly executed skill- based routines and is symptomatic of an overloaded WM.

Where a task is well defined, the problems of management are typically associated with the rates at which inputs can be received, processed and responded to. The number of tasks processed per unit time does become an issue as capacity limits are reached. However, ill-defined tasks are associated with a different range of problems. Unfamiliarity can result in a poor structuring of the situation with a resultant fixation on certain features of the problem. Tunnel vision (focussed attention) limits the range of options considered and cognitive lock-up results in aspects of the problem being treated in sequence as opposed to in order of importance. Poor prioritisation, generally, and incorrect sequencing are also features of coping with an ill-defined task.

We saw in the ATR 42 fuel scenario that individuals exercise agency: they act with a degree of autonomy. In a systems model, SOPs represent a form of control but individuals’ acceptance of constraints on the work process is a variable, not a given. This property of ‘agency’ has a bearing on other aspects of task management. Once the primary goal of work and the associated cognitive requirements have been satisfied then operators can consider other performance modifiers, such as:

  • • perceived risk of failure
  • • current workload
  • • time remaining
  • • perceived social acceptance of task accomplishment

The control of the current task, then, requires the activity to be tracked and evaluated in terms of its validity in relation to the current active goal. Task management requires action sequences to be modified and this increases the workload. The process can fail when the capacity is overloaded or prioritisation is suboptimal. Attempts to control tasks through procedures can fail if individuals elect to work outside of prescribed bounds. Ultimately, though, action at the level of the individual is a stream of continuous decision-making directed at achieving the planned goal.

 
<<   CONTENTS   >>

Related topics