Abstract
Uncertainty plays a vital part in all decision making. The position taken in this dissertation is that the choice of action should correspond to the decision maker’s preferences and to the judgement of the uncertainty involved, consistent with the expected utility model. Due to the decision maker’s lack of information and incapacity to process and transform knowledge about uncertain objects into precise as well as reliable probability estimates, it seems reasonable to allow for imprecise measures, so as to perform a useful analysis. The problem with such analyses is that, the more (less) imprecise information allowed for, the less (more) likely that one optimal action can be identified. To deal with this dilemma of usefulness, a new indifference method is suggested and analyzed in the study. Chapter 2 provides a definition of the choice situation under consideration, and a treatment of the concepts of utility and personal probability. The chapter concludes with a review of methods for analysis of dominance given imprecise probability information. Chapter 3 is concerned with defining the indifference method and discussing its characteristics. Chapter 4 contains an investigation of the confidence in the type of probability estimates required by the indifference method relative to other types of probability estimates. The investigation includes theoretical propositions as well as data from an empirical study of professional probability assessors. Chapter 5 describes a real-life application of the indifference method to the problem of determining the size of magazine editions. The model and methods suggested and tested for solving the problem are discussed with respect to the participants’ perception of the usefulness of the approach. Chapter 6 provides a summary of the report and a few preliminary ideas about future research