Tony O'Hagan - Academic pages

Short course: Introduction to Elicitation

This course provides a basic grounding in the theory and methods for eliciting expert knowledge in probabilistic form. No background knowledge is assumed beyond very basic appreciation of statistical methods.

Learning Objectives

At the end of the course, participants should:

• appreciate the wide range of contexts in which elicitation is found to be valuable;
• understand the subjective interpretation of probability that is important in most practical elicitations, and be able to assess simple subjective probabilities;
• be aware of the importance of psychology in elicitation, and know some of the most common sources of potential biases;
• be conversant with the SHELF system for organising and facilitating elicitation;
• understand some alternative approaches to elicitation using several experts and how this is dealt with in SHELF;
• appreciate the importance of structuring to create independence between two or more uncertain quantities, and the simple and hierarchical methods of structuring;
• be aware of the complexities in more advanced topics such as multivariate elicitation, scoring rules and uncertainty/imprecision in elicitation.

Synopsis

1. What, why, who, where, how? What is elicitation? Why do it? Who should be involved? Where and how should it be organised?
2. Psychology. Judgement and decision making. Bad judgements. Towards better judgements. Limitations of the psychology literature.
3. One distribution, one expert. Eliciting a distribution. A first look at SHELF.
4. Multiple experts. Mathematical and behavioural aggregation. Pooling methods. SHELF again.
5. Multivariate elicitation. Joint distributions. Fitting and copulas. Structuring.
6. Other topics. Topics relevant to the course hosts. This session changes each time the course is given.