Biased experts, biased decision makers: How to optimize the quality of judgments in Decision Analysis?
Gilberto Montibeller, Prof. of Management Science, Loughborough University, UK
July 17, 2019 (Joint SDP/DAS Webinar Invited Talk )
Abstract: Decision analysis models rely heavily on judgments of experts and stakeholders, for instance to estimate the parameters such as probabilities of events, risk tolerance and value trade-offs. In the 1970s, psychologists including Kahneman & Tversky demonstrated that such judgments can be systematically biased. Since then, behavioral decision researchers have identified hundreds of cognitive biases that affect judgments.
Moreover, we now understand better how experts and decision makers may also suffer from motivational biases, which can drive selective attention and generate distorted assessments. In addition to these individual biases, social psychologists found that group biases and dysfunctional group behavior may also affect judgments.
These biases may cause serious detrimental effects on the quality of our models. However, while the earliest of these findings are reflected in decision analysis techniques described in textbooks, the current body of behavioral literature is largely disconnected from the prescriptive practice of decision analysis.
Decision analysts are thus confronted with tough elicitation tasks without clear guidance: Which biases really matter? Which bias should we target? In which modeling steps do they creep in? And which debiasing strategy should we employ?
In this webinar you will learn what are the most relevant biases for decision analysis. You will explore the use of debiasing strategies against biases for every modeling task requiring a judgment. You will learn how we can debias experts and decision makers in practice.
The webinar is based on an extensive evidence-based review of the subject and on my experience of helping many global health organizations with decision analysis. This talk will help you achieve our goal: optimize the quality of human judgments in our models.
Click on the file below to hear a sample of the presentation.