Mini-review: ‘Judgment and Decision Making as a Skill’
A new book from Cambridge University Press describes the impetus of the forthcoming Rationality Group in its title: Judgment and Decision Making as a Skill. It begins:
Our scientific understanding of human judgment and decision making (JDM) has grown considerably over the past 60 years in terms of the normative benchmarks… by which we assess performance, the descriptive models we use to describe JDM, and the prescriptive models we offer to improve JDM...
...[But] how do we learn to make good decisions? How can we improve or aid our decision making? Fortunately, there is an emerging body of work that is interested in long-term and short-term changes in JDM skills… There is research on the acquisition of expertise in JDM, and training and aiding of JDM. Researchers more interested in short-term changes have begun to study learning of JDM tasks...
[We] introduce a new conception of JDM, seeing it as a dynamic skill rather than a static capacity...
Chapters 1 and 2 survey the evolution and neurobiology of JDM, while the chapters 3-5 discuss JDM in young children, adolescents, and the aged. Chapters 6-10 were the most interesting to me, because they concern the learning and improving of JDM skills.
In particular, chapter 7 discusses the use of causal Bayes nets to model JDM processes and thereby make better-informed choices among possible debiasing interventions, and chapter 8 discusses JDM in the context of skill-learning (from feedback). Chapter 9 reviews the ways in which JDM can be improved simply by communicating and representing information in particular ways. Chapter 10 reviews “procedures or devices that are intended to improve the quality of people’s decisions.”
Chapter 11 contains personal reflections on JDM as a skill from nine past presidents of the Society for Judgment and Decision Making.
Overall, the book is a handy collection of review articles on JDM (what LW calls epistemic and instrumental rationality) written from a useful perspective. But it is not as useful as Stanovich’s Rationality and the Reflective Mind, and I anticipate it being less useful than the forthcoming Oxford Handbook of Thinking and Reasoning.
The nearby Stanford U is one of the early adopters of teaching decision-making science. Using Bayes Networks (actually Influence Diagrams, which is an extension that includes utility nodes) is largely due to Ron Howard who is responsible for much of the curriculum there. Many of the courses are online and can ba audited for a smaller fee. Vaniver’s sequence covered much of the introductory material.