On the Relevance of Behavioral Decision Research for Decision Analysis

Abstract

In 1954 and 1961 Ward Edwards published two seminal articles that created behavioral decision research as a new field of study in psychology (Edwards, 1954(Edwards, , 1961. The topics of this research include how people make decisions and how these decisions can be improved with tools and training. Behavioral decision research has been conducted in two distinct paradigms: the cognitive illusions paradigm and the engineering psychology paradigm. Both are, in different ways, relevant to decision analysis.The cognitive illusions paradigm evolved from the work by Tversky and Kahneman (see Kahneman, Slovic and Tversky, 1982;. It is concerned with describing how people make judgments about uncertainties and values and how these judgments deviate from probability theory and expected utility theory. The term cognitive illusion was used by Ward Edwards (see von Winterfeldt and Edwards, 1986, chapter 13) to draw attention to the similarity of this paradigm with the study of perceptual illusions. J. Shanteau et al. (eds.), Decision Science and Technology © Kluwer Academic Publishers 1999 134Research within the cognitive illusions paradigm typically involves four steps:1. Identify a normative model for a judgment or decision making task (e.g., the multiplication law of probability theory); 2. Experimentally demonstrate the deviations from the normative model (e.g., the conjunction fallacy); 3. Develop an alternative model or explanation (e.g., representativeness); 4. Test the alternative models or explanations.This paradigm has stimulated an extraordinary amount of experimentation in many scientific areas including cognitive psychology, social psychology, economics, management science, and medicine. It has led the way to new descriptive theories of how people judge uncertainties and uncertain prospects.The engineering psychology paradigm has a long-standing tradition in the human factors and human performance sub-fields of psychology. It is concerned with improving people's judgments and decisions. Researchers using this paradigm consider deviations from normative models as a challenge to be overcome by training and by developing tools to improve human performance.Research within this paradigm follows the typical engineering approach of developing tools for improving performance, to test these tools in experimental settings, and to implement the successful tools. Typically, it involves the following steps:1. Identify a normative model for a judgment or decision making task (e.g., Bayes' theorem for revising probability judgments); 2. Identify implementation problems (e.g., conservatism or ignoring base rates); 3. Develop tools to overcome the implementation problems (e.g., decomposition of the task into judgments of prior odds and likelihood ratios); 4. Test the tools and implement successful ones.Research within the engineering psychology paradigm has produced useful tools for improving judgment and decision making.It has provided experimental support for the value of decision analysis and it tested the models and tools of...

Cognitive and Motivational Biases in Decision and Risk Analysis

“…show that the elicitation method (fixed value vs. fixed probability) influences the results, mainly due to the anchoring bias . The same bias also occurs in eliciting probability distributions when the expert uses a small set of data to make the estimates and does not include alternative scenarios …”

“…We define motivational biases as those in which judgments are influenced by the desirability or undesirability of events, consequences, outcomes, or choices (see also Kunda, von Winterfeldt, and Molden and Higgins). An example of a motivational bias is the deliberate attempt of experts to provide optimistic forecasts for a preferred action or outcome.…”

“…In addition to promoting precise definitions of events and hypotheses, it is common in this step to use group elicitation and counterfactuals to stimulate creativity and try to reduce the omission and availability biases. The evidence about the quality of group interaction indicates that nominal groups (where members generate ideas in isolation) outperform traditional brainstorming groups, in terms of both heterogeneity and quality of ideas .…”

“…There is a strong influence of the scaling biases on probability elicitation, for example, the use of a linearly spaced or logarithmic spaced variable is likely to influence the results . Seaver et al .…”

Behavioural decision theory for multi-criteria decision analysis: a guided tour

“…Unlike these books, this 'guided tour' explicitly focusses on the challenges facing the practising analyst. Excellent review papers exist, in particular those of Payne et al (1999) andvon Winterfeldt (1999). Our exclusive focus on MCDA and coverage of recent developments in BDT differentiate us from these, and in any case, our perspective differs so our coverage only partially overlaps with that of these works.…”