There seems little doubt that decision making plays an important part in coaches’ everyday practice and is a significant component of coaching expertise. It might be argued, however, that coach education has yet to pay sufficient attention to this aspect of the coach’s skills set. This chapter makes the case for the centrality of decision making in coaches’ practice and offers the Naturalistic Decision Making (NDM) paradigm (Lipshitz et al 2001) as an appropriate mechanism for describing and explaining how coaches make decisions. As subsequent sections will demonstrate, NDM is a (relatively) recent research paradigm in psychology that focuses on the decision making of proficient and experienced practitioners. The more traditional Judgement/Decision Making (J/DM), based on the rational assessment of alternatives (Teigen 1996), did not appear to offer an adequate mechanism for describing and understanding the apparently intuitive decision making of these practitioners. NDM (Cannon et al., 1996, Flin et al., 1997, Zsambok and Klein, 1997, Lipshitz et al., 2001 and Montgomery et al., 2004) offers a number of alternative models of decision making, the best known of which is Klein’s (1998) Recognition Primed Decision model. This model illustrates the central theme of NDM, which is that for proficient practitioners the decision to act ‘emerges’ from a largely sub-conscious process in which the decision is generated from previous experience and is based on a reading and recognition of the situation.

Although the literature referred to will confirm the recognition of decision making as an element of sports coaches’ expertise, it will also demonstrate that there has been very limited exploration, in theory or concept, of the mechanisms through which the coaches’ cognitions enable this decision making. Therefore, this chapter offers a research agenda, and provides an account of how the NDM paradigm could provide insights that resonate with coaching practice.

Practising coaches will have no problem with an assumption that much of their functioning as coaches involves taking decisions. Indeed, if we conceive of the coach’s behaviour, in all but its most instinctive or reactive forms, as being the result of an ‘action decision’, it might be argued that decision making assumes a pre-eminent place in the coach’s expertise. Of course, conceptions are never as straightforward as this, and it is necessary to distinguish between deliberative and non-deliberative decision making. Deliberative decisions involve those aspects of practice in which the coach has time and space to consider options and weigh the relevant evidence to decide upon the most appropriate course of action. This will characterise much of forward planning, but also aspects of team selection, reflecting on progress, considering tactical options, or dealing with some relationship issues. On the other hand, non-deliberative decision making is perhaps most evident in game or competition situations in which the coach has to take decisions under considerable time pressure (perhaps thought of as ‘reactive’). This may be too ‘black and white’ for the messiness of coaching practice, in which much of the coach’s decision taking can be categorised as semi-deliberative; that is, there is an element of time pressure, but no suggestion that the coach’s behaviour is (or should be!) completely without conscious deliberation. It is not too much of a jump to characterise the greater part of the coach’s management of interventions (the training session), competition, and interpersonal interactions as semi-deliberative.

The reason for introducing these distinctions at this early stage is that NDM will be shown to contribute most to the coach’s decision taking in semi-deliberative and non-deliberative circumstances. It will also be obvious to students of sports coaching that the literature will show, or at least conceive, that much of the coach’s action decision taking appears intuitive and draws upon personal knowledge in ways that are described as tacit or implicit. It will be necessary, therefore, to demonstrate throughout this chapter that NDM offers insights into such behaviour, and can make a contribution to both research and coach education.

The chapter begins with a brief account of the decision-making circumstances in which NDM provides the most apt explanation for decision-making behaviour, and demonstrates why this is particularly appropriate for sports coaching. This is followed by a more in-depth examination of NDM, which illustrates how some of its findings resonate with semi-deliberative decision making in coaching practice. There are two focused sections. The first provides an example of ‘story telling’ as an exemplar of NDM-related analysis of coaches’ narratives; the second examines intuition and the role that this plays in NDM accounts of decision making. The chapter concludes with an identification of the implications for understanding sports coaching expertise and designing coach education.

An acknowledgement that the sports coaching practice of expert coaches is a predominantly cognitive activity is relatively recent (Côté, 1998 and Vergeer and Lyle, 2007), although research into coaches’ cognitive organisation is clearly not yet a coherent field (Gilbert & Trudel 2004). Studies have focused on general decision processes (e.g. Jones et al., 1995 and Jones et al., 1997) or specific coaching problems (e.g. Vergeer & Hogg 1999). This latter study is an example of a policy-capturing approach in which the focus is on identifying the ‘rules’ used by coaches to make decisions in given circumstances. There have been notable attempts to study coaches’ knowledge structures in relation to coaches’ practice (e.g. Saury & Durand 1998) but coaches’ decision making has received limited attention. To some extent this is surprising given that recent literature identifies decision making as the core of coaches’ expertise. For example, Abraham et al (2006) suggest that coaching is ‘fundamentally a decision-making process’ (p 549), although they fail to elaborate on what this means for coaching expertise. Nash and Collins (2006) reinforce the point, stating that decision making is ‘one of the key functions that define a coach’ (p 466), adding that knowledge gained from everyday (practice) is ‘tacit’ and often ‘unarticulated’ by the coach.

A simple and workable definition of decision making is provided from within the NDM paradigm: a decision is ‘committing oneself to a course of action’ (Lipshitz et al 2001). It may therefore be more profitable if we work on the assumption that coaching requires decision making, and as with other professions it constitutes the mark of the professional (Beckett 1996). At the same time, it is necessary to keep in mind that the range of decisions will stretch from the automatic and unconscious; the routine, repetitive decisions about which there is no conflict; decisions with a choice of alternatives (with impact on achieving goals); to fluid, real-life decisions in which there is no clarity about the alternatives or the basis on which the decision might be taken – perhaps better thought of as problem solving (Svenson 1996). It does not take too much imagination to visualise this as part of coaching practice. The useful categorisation of deliberative, semi-deliberative, and non-deliberative helps to bring some order to the mix of routine training interventions, coach–performer interactions, planning and management activities, crisis interventions, competition management, and so on. The ‘pressured’ action in response to a performer’s reaction to the coach and the control of the competition situation with key momentum-turning decision points are often visible manifestations of the coach’s skills (or lack of skills!). Nevertheless, the cumulative, and perhaps more reflective, decisions about managing the progress of the training interventions may be even more crucial to achieving success.

Lyle (2002) suggests that the aim of the coach is to ‘manage uncertainty and retain control’ (p 137), while acknowledging the factors that make this difficult. However, there is limited value in identifying the uncertainty that undoubtedly characterises coaching practice without working towards an explanation for how the coach copes with this. It is also worth pointing out that ‘control’ does not refer to performer behaviour but to maintaining focus on the individual and group targets and goals that have been agreed. Lyle (2002 p137) offers some hypotheses about decisions taken. These highlight the balance between impact and certainty in making decisions, the ability to reduce option choices, the influence of personal biases and preferences, and the use of heuristics that optimise rather than maximise outcomes. He cites Lipshitz and Bar-Ilan (1996) who suggest that we should diagnose (recognise the need for action) as early as possible, delay decision taking, and select actions that do not completely restrict our ‘room for subsequent manoeuvre’. For coaches reading this text this advice is likely to resonate with their experience and practice.

The purpose of the chapter is to draw attention to the extent to which NDM offers a plausible explanation for coaches’ decision making. Teigen (1996) distinguishes between the traditional experimental, laboratory-based Judgement/Decision Making (J/DM) research with structured, contrived problems, and NDM, which is ill-structured, ‘messy’ and untidy, and with fewer ‘givens’. The question therefore, is whether sports coaching offers the sort of decision-making environment within which NDM analyses and explanations might be profitably applied. Randel et al (1996 p 579) describe the contrast between NDM and more traditional research problems. They characterise NDM contexts as: dealing with ill-structured problems; suited to uncertain dynamic environments; operating best with shifting, ill-defined or competing goals; process-orientated; time-constrained problems; having outcomes with high stakes; and having multiple actors involved. The ‘high stakes’ element is derived from the kind of outcome consequences that accompany military fire control officers, firefighters and air traffic controllers. Although there may be personal, economic and career consequences from coaches’ decisions, the day-to-day practice of the sports coach does not satisfy this criterion. However, recent descriptions of the coaching environment (e.g. Bowes and Jones, 2006 and Cushion, 2007) match extremely well with all of the other characteristics.

Sports coaching involves decision making, particularly in semi-deliberative practice, but this is not of the ‘choice of options’ type. Not only is it rare to be able to demonstrate that an action decision was ‘correct’, but practice generally consists of a series of inter-dependent decisions. Therefore, decision taking is not about option choices (in the language of Judgement/Decision Making psychology) but about coming to the most appropriate decision on an on-going basis. The decision making of intervention and progression by coaches should not be thought of as necessarily problematic (although the literature speaks of problem-solution linkages), but merely as the minute-by-minute job of the coach in managing intervention, interaction and progression in complex circumstances.

Naturalistic Decision Making (NDM) has established itself as a ‘force’ within psychological research over a period of 10 or 15 years (Cannon et al., 1996, Flin et al., 1997, Zsambok and Klein, 1997, Lipshitz et al., 2001 and Montgomery et al., 2004). NDM ‘is an attempt to understand how people make decisions in real world contexts that are meaningful and familiar to them’ (Lipshitz et al 2001 p 332). NDM emerged in response to an awareness that decision makers who could be described as proficient, that is, had considerable knowledge and experience in a professional domain, appeared not to use a process that considered alternative options, and also appeared to operate at a tacit or sub-conscious level. This decision-making process is most useful in circumstances characterised by complexity, uncertainty, goal conflict, and time constraints. NDM describes decision making where there is a dynamic interplay of experience/knowledge, a high level of complexity, and the environment, and where the generalisability of laboratory-based research would be questionable (Currey & Botti 2003). In these circumstances, the limitations of human cognition become evident, and resultant actions emerge as appropriate rather than being measured by their distance from the ‘correct’ solution (Meso et al 2002).

Lipshitz et al (2001) in their overview of NDM point to several differences from the traditional J/DM. NDM is about matching rather than choosing, is process-orientated rather than output-driven, and is context-bound rather than context-free (p 334). The centre of focus is the experienced decision maker. However, the field setting is important because its constraints and affordances shape the decisions taken. NDM was a response to the realisation that expert decision makers who cope with complex, novel and dynamic situations do not employ a (fully) rational choice approach to their decision making. An important feature of NDM is the emphasis on the process of taking decisions, rather than the action selected. In general, NDM models are variants of ‘situation-action matching’. In other words, proficient decision makers take action because it is appropriate for the situation, not because it is better than the alternatives. Appropriate actions ‘emerge’ because of their compatibility to situational needs. The process is based on an expert knowledge structure that is domain- and context-specific, and in which attention is paid to the semantic content (meaning) inherent in the situation.

There are a number of descriptive models within NDM. The best known of these is the Recognition-Primed Decision making (RPD) model (Klein 1998). This model in essence describes pattern recognition of a situation leading to a known course of action. More elaborate versions deal with enhanced diagnosis and a reconsideration of action decisions. Klein, 1993 and Klein, 1998 Recognition-Primed Decision model is a development of the immediate reaction to the perception of the problem space. Klein stresses the situational assessment that allows the experienced worker to immediately generate the most appropriate response based on previous experience. The individual adopts the solution or action associated with the ‘recognition’, which obviates the need for consideration of multiple alternatives. This reinforces the ‘recipe-led’ executive command response to recognised patterns.

Time pressure is a common feature of decision contexts in which NDM has been used to explain behaviour. One element of this is the opportunity–cost equation. This contrasts the limitations of ‘deciding too soon’ (but not being sure of the situation) and delaying choice (which gives a lower return). Payne et al (1996) suggest three ways to react to time pressure: accelerate the processing (spend less time deciding), become more selective, or change one’s decision strategies. This is made easier by using simple and fast rules of thumb (heuristics). For example, Payne and colleagues suggest that when under time pressure the decision taker should identify and use the important attributes in the first instance, and consider options only on those attributes. These speedy search heuristics allow the decision maker to recognise common domain-specific patterns (Devine & Kozlowski 1995). In other words, experts can ‘decompose’ complex and ill-structured situations and convert to more recognisable and manageable attributes. This is the ‘is able to read situations and people’ that we may recognise in those who cope well with the need to deconstruct unusual circumstances and events.

Many of the NDM explanations deal with the issue of risk or threat (firefighting, nursing care, command and control, air traffic control). This implies that physical harm may follow, or the consequences of an inappropriate decision may lead to serious consequences. For the sports coach the consequences are much less dire, and may involve a threat to performance objectives or personal relationships. The ‘reading’ of threat is one of the learned abilities for which NDM offers an explanation. This ‘reading’ may be problematised by uncertainty or time pressures. In such circumstances a default option or weak solution may be employed (Orasanu 1997). In practice, this often means that coaches respond with a generalised action that has wide coverage (address many circumstances), but only attends to the issue when that default response does not appear to ‘work’. An example might be a coach who makes a substitution because of a lack of momentum, but considers more specific tactical changes once the problem persists. One of the hypotheses that deserves attention is that novice coaches interpret ‘threat’ too soon (do not allow variation, and respond to too many issues), whereas the more expert coach may be willing to allow the circumstances to settle down or accept less ‘smooth’ progress. This hints at the education and training process within which Lipshitz and Strauss (1997) suggest that ‘programmes should aim at teaching novices and medical performers the strategies and tactics that are used by experienced decision makers in the same domain’ (p 160).

Randel et al (1996) used situational awareness to examine the differences between expert and novice coaches. Their findings were that experts put greater emphasis on the situational analysis rather than the decision taking. This may be explained by the expert having had more opportunities to benefit from feedback and therefore a more extensive repertoire of ‘solutions’. As a result, the solution or required action becomes ‘obvious’, providing the reading of the situation has been accurate. Randel and colleagues also found that experts had the ability to recall patterns visually, although these had to be meaningful and domain-specific. Their conclusions were that experts would spend time on understanding and assessing the situation, rather than evaluating the best course of action. The distinctions between novices and experts were reflective of the experts’ integration of cues, knowledge and imagery, rather than each individual element. In relation to the heuristics or mechanisms for prioritising options, they suggest that ‘experts do not have more or different rules of thumb than novices but they do show a better knowledge of how to apply the rules’ (1996 p 594). Randal et al also reinforce the concept of modelling, ‘expertise appears to take the form of a complex model of potential situations. For experts, rules of thumb are a short-lived expression for the cause and effect relations of a complex model that they have internalised’ (1996 p 595).

Kaempf et al (1996) also emphasised situational awareness. Individuals used ‘feature matching’ to recognise or categorise events that matched ones they had experienced previously. These ‘key attractors’ or cues in the display will be domain-specific, and perhaps even difficult to verbalise. If the situation or event is novel or complex, the decision maker will ‘story build’. Story building refers to a process of analysis and interpretation based on previous experiences. Another hypothesis is that the expert coach can identify the most appropriate ‘cues’ and ignore the irrelevant.

Before moving to a couple of NDM-related themes, it is important to emphasise the level of cognitive organisation required, and perhaps why NDM is a mark of the domain expert. The NDM explanation for decision-making behaviour is predicated on accessing ‘organised’ knowledge. The knowledge is stored in a mixture of knowledge ‘about’ and knowledge ‘how to’. It is also worth pointing out that the expert’s ability to ‘mentally simulate a course of action and anticipate how it will play out’ (Lipshitz et al 2001 p 336) is a form of ‘propositional’ knowledge; in other words, ‘if I do this … that will happen’. This form of knowledge is not necessarily based on having previously taken this action in exactly similar circumstances, but the expert is able to generalise (forward reason) from previous examples, and apply the mental simulation to the newly experienced context. What emerges is a family of terms (mental simulation, reflective anticipation) that suggest a ‘feedforwardness’. The sports coach who introduces a new drill, who has to deliver selection decisions, or who loses a player to injury during a competition has experienced a similar situation previously and is able to ‘image’ or predict what might happen and has already generated an appropriate solution. Beckett (1996) p 140 suggests that ‘an anticipative conversation with our practices’ is a good descriptor of this aspect of the professional’s ability to cope with dynamic ill-structured contexts. The expert is constantly thinking ahead about the potential consequences of actions taken, and this is continuously monitored as the circumstances unfold and the reactions and responses of the other actors become clearer. This is largely conducted in a selective fashion and at a subconscious level, acting as a form of ‘hypothesis narrowing’; the coach becoming gradually more certain that the situation is being understood. Of course, coaching practice is made up of a myriad of interpersonal interactions, performance-related decisions, changing contexts, and so on. These are happening in a continuous stream of sub-episodes that are at different stages of certainty and outcome. It is hardly surprising, therefore, that coaching is described using the terms messy, complex, dynamic, and so on.

From the literature we can point to a series of hypotheses that might populate a coaching research agenda. We might reasonably speculate that expert coaches will be better able to ‘model the future’ because of more accurate predictions. The coach’s behaviour will continue to seem ‘intuitive’ – the actions taken have already been weighed and judged (albeit rapidly and often sub-consciously) to be appropriate; actions follow as the need to respond becomes pressing. It is worth recalling that not all decisions are non-deliberative. For the most part there will be an element of deliberation about the coach’s behaviour, although the sheer scale of the potential interactions and decisions may help us to appreciate why the ‘rational analysis’ approach seems implausible. The need to ‘cope’ with the potential range of action decisions also explains why experts use heuristics – rules of thumb that help to simplify the problem of option choices. Identifying the use of heuristics by our best coaches is another task for the researcher.

The apparently intuitive, often sub-conscious, reasoning of the expert is dependent on accessing context-referenced frames of knowledge; that is, the knowledge is domain- and context-specific. It is developed from and through experience, and is meaningful to the individual (for example the coach) only in that context. This specificity of knowledge frames is very significant for the construction of learning programmes in coach education, and questions the genericism of problem-solving capabilities. The speedy form of expert decision making is contrasted with the more deliberative, analytical and explicit behaviour of the novice (Boreham 1994). The expert’s cognitive style is largely ‘routinised’ and under schematic frame-recognition control. Evans (1989) describes a schema as a ‘knowledge structure which is induced or learned from experience, contains a cluster of related declarative and procedural knowledge, and is sensitive to the domain and context of the current focus of cognitive activity’ (p 84). These schemata are the frameworks through which we represent the world to ourselves. A good deal of work has been done on clinical decision making by experts. For example, Schmidt et al (1990) describe how doctors move from memory-based knowledge structures to memories of specific illnesses and then to memories of specific patients as they become more expert. This reinforces the move to implicit behaviour by stressing the use of ‘scripts’ based on particular instances. Scripts are a particular kind of knowledge structure that predicts sequences of events, including likely solutions to problems. Experts recognise the similarity of cases to previous examples and apply highly idiosyncratic ‘recipe’ solutions.

The reader will be able to recognise the parallels in the coach’s practice. The recognition of instances of interaction with performers, training-ground problems, and competition patterns in relation to previous typical cases or specific cases is one way that the coach can enhance situation awareness and reduce the need for deliberation and ‘trawling through options’. We can hypothesise that the coach will apply actions that address the case, and perhaps ‘attend’ to the peculiarities of the situation only if it becomes apparent that the action is inappropriate. Once again the need for experience and a repertoire of ‘cases’ is highlighted.

In the introduction to this chapter it was suggested that the novelty inherent in the application of the NDM paradigm to sports coaching would have the effect of creating a potential research agenda, and opportunities have been taken throughout to suggest appropriate questions. Vergeer and Lyle (2009) illustrate the insights that might be obtained from such research. In a study of the decision-making policies and practices of more and less-experienced gymnastics coaches, there were differences in both the amount and value of information attended to by coaches with differing levels of experience. Experienced coaches showed evidence of an extensive knowledge and a capacity to weigh a greater range of factors, although they reduced uncertainty by taking earlier decisions about injured gymnasts’ participation. The coaches’ greater experience appeared to allow them to frame the problem in a broader context. The less-experienced coaches appeared to have less cognitive complexity and more ‘surface’, rather than performance-related, considerations, perhaps suggesting that they attended to a different set of cues. Another obvious focus of attention for researchers is the nature of the ‘mental models’ that are becoming acknowledged as an essential part of the coach’s cognitive organisation. Mental models were recognised by Côté et al., 1995 and Lyle, 2002 simulation model fulfils the descriptor of the mental simulation of future action and how it might unfold given by Lipshitz et al (2001).

There is clearly substantial benefit from applying NDM methodologies and themes to the coaching context, and perhaps reinforcing the relevance of the paradigm for understanding coaches’ behaviour. However, the coaching context also offers something to NDM research. In addition to a serial multi-decision context, the contested aspect of sport creates circumstances in which competition-coaching behaviour, for example, is being actively opposed by another coach. The ‘winning and losing’ element of sport participation, performers’ reactions to their (often publicly evaluated) progress and performance, the coaches (in many instances) familiarity with the performers, and the physical and psychological competition between athletes create a working context in which emotion and emotional responses play a significant role. It seems likely that decision making will be impacted by the emotion and personal meaning attached to actions.

A study into non-deliberative decision making by expert coaches (reported in Lyle 1999) identified a Slow Interactive Script model as the predominant cognitive mechanism. This suggests that coaches were using a strategy of continual refinement of action and potential action against ‘scripts’, that is, images or models of the situation, its determinants, and likely future outcomes. These scripts are likely to have been very strongly influenced by an accumulation of previous experiences. From these experiences expectations and appropriate behaviours have been shaped into mental models. These in turn, help coaches to recognise what is happening, situate it within existing experience, and choose appropriate actions. These Slow Interactive Scripts are mediated by the coach’s objectives, and the combination of situation, actors, and goals creates a very dynamic and challenging atmosphere. The research was carried out on volleyball coaches during competitive matches and the non-deliberative (time-pressured) nature of the decision making will have exaggerated the fast-moving nature of the context. Coaches in this case appeared to attempt to control uncertainty and the contested nature of the activity through anticipatory modelling of the action.