Exceptional Paper—Probability Encoding in Decision Analysis

Carl Spetzler, Carl-Axel S. Staël von Holstein

Management Science · 1975 · 665 citations · 10 references

Concepts

TL;DR

Probability encoding extracts and quantifies individual judgments about uncertain quantities, a key function in decision analysis. The paper outlines the philosophy and practice of probability encoding at Stanford Research Institute and discusses how to set the encoding process, using sensitivity analyses to identify key state variables. The authors balance modeling and encoding techniques, review biases, present various encoding methods, and recommend a structured interview process with trained interviewers to reduce biases and aid quantification.

Abstract

This paper presents the present philosophy and practice used in probability encoding by the Decision Analysis Group at Stanford Research Institute. Probability encoding, the process of extracting and quantifying individual judgment about uncertain quantities, is one of the major functions required in the performance of decision analysis. The paper discusses the setting of the encoding process, including the use of sensitivity analyses to identify crucial state variables for which extensive encoding procedures are appropriate. The importance of balancing modeling and encoding techniques is emphasized and examples of biases and unconscious modes of judgment are reviewed. A variety of encoding methods are presented and their applicability is discussed. The authors recommend and describe a structured interview process that utilizes a trained interviewer and a number of techniques designed to reduce biases and aid in the quantification of judgment.

References

10