Airmanship, another much-disputed and ill-defined concept, is the apparently effortless ‘reading’ of a situation and then adapting to requirements. What we traditionally call ‘airmanship’ is really knowledge rendered as expertise. Knowledge is stored as scripts and schema (plural schemata). Scripts are canned sets of concepts, facts and rules that guide action in certain situations. I can have a script for eating in a fast food restaurant and another script for eating in a Michelin 3-star restaurant. Both are outlets that provide food but the behaviour expected in each and, therefore, the way of getting fed differs between them. Schemata are mental representations that allow us to control situations and, if necessary, run ‘what if’ scenarios in order to make choices between alternative courses of action. Some authors differentiate between schemata as structural models and ‘frames’ as functional models. From a training perspective, effective performance is based on the development of adequate mental representations of the world. It is a more commonplace to use the term ‘mental model’ in aviation to describe the combination of scripts and schemata. As this is not a psychology textbook, I will use either schemata or mental model to describe the way the brain stores information. To be effective, a mental model must have a number of features, including:
When we consider action in a social, collaborative context, we must add:
This elaboration of the mental model concept points to some potential candidate areas of competence but it also starts to break down the traditional divide between soft and hard, technical and nontechnical skills. So, I cannot construct a model of the dynamic environment without a robust understanding of the procedural framework I am working within. I cannot gauge success if I do not understand how the aircraft system will behave. To control activity, I need an understanding of the information available to me in the workspace and, in order to guarantee effective collaboration, I need to understand what information must be communicated to others. Task-related ‘hard’ skills are enacted in a ‘soft skill’ context in such a way that any attempt to separate the two becomes arbitrary and, probably, counterproductive.
Two key questions concern us: how do mental models develop and how are they used to manage tasks? To answer the first question, schemata are developed as a result of training or after repeated exposure to similar situations. The problem is that repeated exposure will only result in schemata formation if there is congruence between observed events and the stored explanation. Unfortunately, there is a catch: a schema can be incorrect despite there being congruence. We used to think that the sun went around the earth and the penalty for suggesting otherwise was severe. Apparent congruence must be tested. Schemata are modified by enlargement (adding new components) and by amendment (refining the structure) (Marshall, 1995). From an expertise perspective, enlargement means that, through experience, we develop a richer set of constraints that allow us to detect problems faster while amendment is the process by which we fine-tune our constraint set.
Cognitive competence is considered to be a measure of the completeness of an individual’s knowledge about the workspace. Unfortunately, any individuals’ ‘knowledge’ w'ill be incomplete and no two people w'ill hold the same knowledge base about their job. Developments in the field of embodied cognition suggest that, in fact, humans shape ‘the world outside’ to suit the needs of the task at any moment, partly to reduce cognitive demand. This offers further opportunities for individuals to hold different views of the world in that we activate knowledge in response to a perceived need rather than in some absolute sense.
In work-related domains, training serves to refine existing mental models and develop new' task-specific models. Training is supported by company-issued documents, such as manuals. Initial and recurrent training is intended to establish an adequate set of rules and mental models that w ill suffice for the majority of situations. Manuals elaborate on this fundamental information but gaps remain. In a study of private annotations on paper-based aircraft manuals and checklists, Wright, Pocock and Fields (1998) found that the majority of comments were elaborations on the conduct of the procedure or explanations of the rationale behind the action - the ‘how’ and ‘w'hy’ of doing work. Interestingly, the move to electronic checklists, e-flight bags and manuals provided via the internet has removed a source of ‘knowledge’ - textual annotations - from the flight deck.
Operational experience allows us to consolidate mental models through repetition, enlarge them through exposure to diverse situations and amend them through problem solving. Eventually, we w'ill reach a stage w'here we can use mental models to reliably extrapolate to novel situations. This is where we start to get close to expertise. Studies of the differences between experts and novices have found that experts do not necessarily possess more ‘knowledge’ but, instead, it is better organised. Experts also tend to search from the problem to the solution. Problems are decomposed and analysed first before solutions are generated. Novices map solutions onto problems, they engage in pattern matching. Schemata building benefits from constant checking of the behaviour of the w'orld against the expected behaviour, supported by effective diagnosis of any causes of divergence. This process lies at the heart of sense making and is also referred to as metacognition.
Having briefly looked at how mental models are developed, we will now move on to consider at how' they are used. At the level of skill-based behaviour, we know' that pattern matching with the environment triggers schema containing the actions needed for the current task. Rule-based behaviour, equally, can often be automatically triggered by cues. However, where activity is blocked, rule-based behaviour will require some conscious effort in order to select between alternative possible behavioural responses. In this case, environmental cues are compared with stored cues attached to possible responses. In the domain of knowledge-based behaviour, mental models are used to run mental simulations in order to anticipate which behavioural intervention, in the absence of a relevant rule, will most probably allow us to achieve our goal. From the earlier discussion of expert-novice difference, it seems that experts have richer mental models of the world that can be reconfigured in order to run multiple ‘what if scenarios in order to find an explanation for observed events. Novices, on the contrary, possess collections of responses and select the response that is considered most likely to succeed. We see this in ambiguous situations where crews attempt to apply whichever checklist matches the situation most closely.
In a study of the impact of the loss of experienced staff on safety in the UK rail industry, the authors noted a link between experience and safety (Wright, Turner & Antonelli, 2003). The report makes three telling points in the context of this chapter (p. 38):
There is evidence to suggest that experience levels are associated with safety performance in the beginning of the employee’s tenure. Once the employee has passed through an initial learning period of service, experience ceases to be a major factor in accident causation.
[T]he [accident] rate is likely to be mediated by the types of experience the individual has been exposed to, task difficulty and how familiar they are with specific working styles.
Evidence from studies reported indicates that once employees become more experienced, and older, accident rates can rise. It is likely that this is due to individuals making incorrect assumptions about the safety of present situations based on past experiences.
The report offers clues as to how mental models, shaped by experience, might affect operational performance. On the one hand, inadequate models will contribute to increased error rates among junior staff; but on the other hand, increased risk tolerance (or inadequate risk assessment) can result in errors among more experience staff. The performance of train drivers probably illustrates how skilled performance is dependent upon more than just sound cognitive models. Expertise is the ability to enact a solution within the situational constraints while being able to factor risk into plan execution.
In this section, I have explored ideas around knowledge and expertise. The framework developed here (Table 1.2) offers a route of entry into the underpinning knowledge that supports performance. By applying the framework to an airline’s range of operations, more effective training curricula and checking regimes can be developed, requirements that underpin a competence-based approach. Having established the performance expected of crew, the next step is to develop appropriate training interventions. In the next section, I will look at historical developments in training analysis and review some of the current confusion that surrounds the adoption of competence-based approaches to training.