TRADITIONAL APPROACHES to the decision making process have employed analytical models that generate and compare options basea on weighted features. This is often referred to as multiattribute decision making. The deliberate procedures developed by the Armed Forces for operational planning-the Joint Operational Planning and Execution System (JOPES)-represent a systematic application of this approach.' Figure 1 illustrates the basic components in this approach to the decisionmaking process.
Recent studies in real-world settings, including tactical commanders in field environments, have led to a different model of the decision-making process.2 These studies of naturalistic decision making (NDM) have resulted in the development of the Recognition-Primed Decision (RPD) model.3 The RPD model asserts that decision makers draw upon their experience to identify a situation as representative of or analogous to a particular class of problem. This recognition then leads to an appropriate course of action (COA), either directly when prior cases are sufficiently similar, or by adapting previous approaches. The decision maker then evaluates the COA through a process of "mental simulation." Figure 2 illustrates the basic structure of the RPD model both in its simplest version and when the decision maker evaluates options through use of mental models.
In general, RPD reflects the ubiquitous influence of analogy in human perception and problem solving.4 Such analogical thinking has demonstrated both its positive and negative effects at the highest levels of national security decision making.5 The emergence of this new model of decision making has direct implications for issues such as training for command, evaluating the expertise of commanders and designing decision-support systems.' The model suggests markedly different decision-support systems, focusing on accurate situation assessment and case-based reasoning (recalling similar cases) as opposed to the feature-based comparison of options inherent in systems such as JOPES.
However, one must recognize that both the analytic and the recognitional modes of decision making are desirable and, indeed, complementary. In fact, studies of decision making in natural settings have demonstrated that decision makers employ RPD and analytic strategies at different times, depending on the problem situation, their level of experience and other factors.7
Figure 3 compares the strengths and weaknesses of the two strategies. The strengths of each approach essentially mirror the weaknesses of the other. As a result, optimal decision making tends to involve some combination of both modes. For example, in operations planning, initial COAs may be generated by the commander based on analogous situations (RPD-based decision making), and the COAs can then be assessed (by the staff via analytic methods. Conversely, once the staff generates COAs for the commander via analytic methods, recognitional decision making may influence the commander's selection of the ones) to implement. Figure 4 illustrates these "mixed" modes of military planning, indicating the interdependent and complementary nature of the two approaches.
Decision-Making Models and the Levels of War
Factors characterizing naturalistic decisionmaking environments include:
* Time pressure/constraints.
* III-structured problems.
* Uncertain, dynamic environments.
* Shifting, ill-defined or competing goals.
* Multiple event-feedback loops.
* High stakes.
* Knowledge-rich environments.
* High decision complexity.8
Each of these factors is present to varying degrees in military planning at the strategic, operational and tactical levels. In general, the strategic and operational levels certainly allow more time and tend to have greater resources for the planning process and thereby favor analytic planning to a greater degree. However, such factors as the increasing pace of warfare, extended battlespace, ability to mass effects and target strategically, near-instantaneous sharing of situational information and the increasing political sensitivity associated with even tactical actions are causing these levels to merge. …