Intuitive or Deliberative? Decision Process Implications for Space Situational Awareness

What is a good decision? How can good decisions be facilitated? Building SSA and taking the optimal course of action (COA) requires human operators to make reliable, accurate decisions. Achieving consistently high levels of decision making is a challenge because SSA tasking confronts considerable uncertainty, time pressure, and complex interactions. The decision science literature is ripe with debates about the benefits and pitfalls of intuition, responsible for snap decisions, and reasoning which is more deliberative [1]. This paper proposes a model for applying these insights to gain understanding and improve designs for the operational community. Among the most robust findings in the literature is the fact that the manner in which a choice is presented has tremendous power to influence the selection [2,3,4]. If managed properly with training and engineered systems, decision making can be optimized by mitigating the vulnerabilities and augmenting the capabilities of human decision makers. Thus, designers should embrace the responsibility and opportunity to facilitate better decisions in the field. 1. DECISION SCIENCE: INTUTION AND REASONING What is a good decision? How can good decisions be facilitated? These are two ponderous questions that can elude satisfactory resolution in complex domains. Yet answers to these questions are vitally important to defending and promoting the space-based interests of the United States. Building Space Situational Awareness (SSA) and taking the optimal course of action (COA) continually requires human operators to make reliable, accurate decisions. Achieving consistently high levels of decision making is a challenge because SSA tasking confronts considerable uncertainty, time pressure, and complex interactions. Despite significant advances in computing power and artificial intelligence (AI), few critical decisions are made without a human decision maker in the loop. This is because human operators inject a needed diversity of expert knowledge, experience, adaptability, and the authority required to successfully execute SSA tasking. Evaluating the quality of a decision can be elusive (as can bounding a decision); in order to constrain this discussion, “rational choice” will tend toward the economic conceptualizations such as maximizing utility and selection consistency. Expected value computations and maximizing utility have a long history as the de facto decision process [5]. Using that as a measure, one could determine “goodness” of choosing between the higher expected reward for a gamble that presents a 50% chance of winning $100 over a 90% chance of winning $25. Another litmus test for a good decision in the rational world is that choices (and thus decision makers) should be consistent. For example, if choice A is preferred to choice B, and B preferred to C, than A should be preferred over choice C as well. These early foundations fostered a process encouraging decision makers to list the pros and cons of a decision, perhaps use a weighting schema, but one way or another weigh the future benefit (or harm) of making a selection. The result, as sought by the rationalist models, should drive toward higher utility. However, connate in human decision processes are mechanisms working against logical, rational thought. Even experts (in their domain), show vulnerabilities in decision making and the empirical research continually generates findings that transcend expertise, rank, and domain. A stronger understanding of the application of decision science could allow system developers to better design tools and craft decision contexts to facilitate better decisions. This paper will highlight several potential improvement opportunities of particular interest to SSA. A wealth of literature on human decision making discussing rational and irrational outcomes exists, proposing varied prescriptive and descriptive models of judgment and decision making [1,6,7,8]. Though consensus among theorists remains elusive, a mounting body of literature shows that the rational, economic models are more brittle than originally believed, and deliberate listing and evaluation of all options is NOT representative of how many decision are made. Emerging is a substantial body of research has eschewed the rational models in response to the poor fit often encountered between the predictions of utility theories and experimental observations of human behavior. In particular, experts in operational settings seem to select the best COA with limited deliberative thought and consideration of the pros and cons of each possible choice [9,10]. A framework gaining interest lately describes two systems predominantly at work: intuition and reasoning [1,11]. Intuition is fast, automatic, and parallel contrasted with the more effortful, deliberative, and sequential reasoning. However, adhering to a path of empirical findings does little to disambiguate what should be the standard of practice. For example, the merits, quality, and purpose of intuition are continually debated. One body of researchers calls intuition a hallmark of expertise responsible for rapid, optimal decisions in the face of adversity [9]. Others espouse intuition as vulnerability where biases lay decision traps leading to unfavorable choices [4]. Gaining favor in academic labs and popular science discussions, these dual process theories also offer promise for improving decision support systems. Though no one theory seems able to withstand all operational settings in all conditions, hybrid approaches hold promise for addressing the diversity of decision conditions encountered in the field. Many findings are showing sufficient value for information architects or system/interface designers to engineer effective decision support tools. 2. DUAL PROCESSES: ERROR AND SUCCESS Instead of impugning theories, we hope to formulate a useful model for applying these insights to gain understanding and improve designs for the operational community. Our belief is that intuition and reasoning (i.e, deliberation) are both important elements in successful decision making. If managed properly with training and engineered systems, decision making can be optimized by mitigating the vulnerabilities and augmenting the capabilities inherent in human decision processes at the intuitive and deliberative levels. To document this approach, a full-factorial dual process theory is proposed (Fig. 1): Fig. 1. Full factorial dual process decision theory. Commitment to a hybrid model is consistent with a data driven-approach in which empirical results, not dogmatic adherence to a theory, guides selection and application of principles, mechanics, and manipulations for decision support services. Since there is evidence that the intuitive and deliberative systems can produce both successes and errors, these results should be studied and applied to better understand under what circumstances errors and successes are the most likely outcome. Below, we provide examples from the literature, and have recast the inquiries to approach SSA interests.